AD-AO~ 81 WSI4IGTONUNI SEATLEF/B 6/18

BIOLGICAL DAMAGE THRESHOLD INDUCED BY ULTRASI4ORT FUNDAMENTAL, -ETC(U)ILDEC 80 A P BRUCKNER, J M SCHURR, E L CHANG F33615-7B-C-0616

UNCLASSIFIED 61-5269 SAM-TR-80-47 NL

E E EA0 81 WSIGTNI E ATE E7E E E

Report SAM-TR-80-47

I BIOLOGICAL DAMAGE THRESHOLD INDUCED BY ULTRASHORTFUNDAMENTAL, 2ND, AND 4TH HARMONIC LIGHT PULSES

00, FROM A MODE-LOCKED Nd:GLASS LASER

C Adam P. Bruckner, Ph.D.

J. Michael Schurr, Ph.D.Eddie L. Chang, Ph.D. '5 0Aerospace and Energetics Research Program

University of Washington

Seattle, Washington 98195

December 1980

Final Report for Period April 1978 - January 1980

[Approved for public release; distribution unlimited.

Prepared forUSAF SCHOOL OF AEROSPACE MEDICINEAerospace Medical Division (AFSC)Brooks Air Force Base, Texas 78235 , ",0I

I~~~ 7)~ O

NOTICES

This final report was submitted by Aerospace and Energetics ResearchProgram, University of Washington, Seattle, Washington, under contractF33615-78-C-0616, job order 7757-02-55, with the USAF School of AerospaceMedicine, Aerospace Medical Division, AFSC, Brooks Air Force Base, Texas.Dr. Taboada (USAFSAM/RZL) was the Laboratory Project Scientist-in-Charge.

When U.S. Government drawings, specifications, or other data are usedfor any purpose other than a definitely related Government procurement opera-tion, the Government thereby incurs no responsibility nor any obligation what-soever; and the fact that the Government may have formulated, furnished, or inany way supplied the said drawings, specifications, or other data is not to beregarded by implication or otherwise, as in any manner licensing the holder orany other person or corporation, or conveying any rights or permission to man-ufacture, use, or sell any patented invention that may in any way be relatedthereto.

The animals involved in this study were procured, maintained, and usedin accordance with the Animal Welfare Act of 1970 and the "Guide for the Careand Use of Laboratory Animals" prepared by the Institute of Laboratory AnimalResources - National Research Council.

This report has been reviewed by the Office of Public Affairs (PA) andis releasable to the National Technical Information Service (NTIS). At NTIS,it will be available to the general public, including foreign nations.

This technical report has been reviewed and is approved for publication.

'- o --I

JOHN TABOADA, Ph.D. /AONN E. PICKERING, M.S.Project Scientist Chief, Radiation Sciences Division

ROY L. DEHARTColonel, USAF, MCCommander

UNCLASSIFIEDS-CURITY CL A SSIFIC ATICN OF r., 'A I L,,

. REPORT DOCUMENTATION PAGE i B (7!AI-.H.":.

-1p -0 -+MiER IZ GC)VI A(CLFE',SION NO' F' CPIEN"S -A' A

SAM4TR-80-47 --.-4. T I T I_~ r F (-0 11.) T'.I1a-&-F " HPr -Pf O ~f', I~ R E

BIOLOGICALDAMAGE THRESHOLD INDUCED BY ULTRASHORT Final Report.. FUNDAMENTAL, 2ND, AND 4TH HARMONIC LIGHT PULSES April 1978 --January 1980.

FROM A MODE-LOCKED Nd:GLASS LASER. 6, F.P'"F6MING ORG EP1o P NUMBER

/. / 61-52697 AUTHOR . - ...T 40 "".' " G A-'-.-" - N

Adam P., Brqckner.. Ph.D.J. . Michael'Schurr, Ph.D. / ) FTF1'-7r-C-OlF.Eddie L.!Chang , Ph.D. __"_ _

1- Pt-RORMfNG ORGANIZATION NAMF A14 ACCR F;, F'I 5A Z.4NARE A 5 WVIDR tN - .L

Aerospace and Energetics Research Program I6 _/.University of WashingtonSeattle, Washington 98195- __ _______

I . CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE

USAF School of Aerospace Medicine (RZL) 1/ December 1980Aerospace Medical Division (AFSC) - - NUMER OF PAGES

Brooks Air Force Base, Texas 78235 8214. MONITORING AGENCY NAME & ADDRESS(Il different from Controlling Office) 1S SECURITY CLASS. tof this repc,rt

Unclassified15a. DECLASSIFICATION DOWNGRADING

SCHEDULE

16. DISTRIBUTION STATEMENT (of this Report)

Approved for public release; distribution unlimited.

17 DISTRIBUTION STATEMENT (of the ahstract entered ir Block 20, If different fror.' Report)

18 SUPPLEMENTARY NOTES

Picosecond biological damage Laser-induced corneal damage, 265 nm/Nd:Glass laser, mode-locked Macac3 fascicularis monkeyBiological macromolecules/DNA, Ocular damage thresholdshilii layer vesiles

Laser-induced retinal damage, 530 nm/ Membrane disruptionMacaca fascicularis monkey

2) Ad -- ACT (C-rrrlnrir orn revere" srI,. If e .' end i PerltIvtrytrhlrr k n-mt,er

-Selected biological macromolecules and 20 Macaca fascicularis monkey eyes wereirradiated with ultrashort pulses of light of various wavelengths derived froma mode-locked Nd:Glass laser in order to determine threshold damage mechanisms.Macromolecules such as calf-thymus DNA, dipalmitoyl phosphatidyl cholinevesicles, and egg-yolk lecithin vesicles, which are similar to the constituentsof living cells that may be susceptible to damage, were irradiated with singlepicosecond pulses and entire mode-locked pulse trains of 1060-nm and 530-nm

D D J A N 7 1 4 7 3 U N C L A S S I F L E D _ _ _ _

SECURITY CLASSIFICATION F piS

UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE(R o. Data Enfored)

20. ABSTRACT (Continued)

light. DNA was not damaged at any energy density available including levelssufficient to cause dielectric breakdown in water. Experimental studies ofelectromagnetic stress-induced birefringence in DNA base-pairs were also cdrriedout in an attempt to establish a lower limit on the restraining forces governingtilting of the DNA bases with respect to the helix axis. The irradiationexperiments at 1060 nni with the bilipid layer vesicles indicated a dama]tthreshold of -600 mJ/cm' for entire pulse trains of -100 pulses and -9 mJ;cm

2

for single pulses# as determined by dynamic light scattering. In both themultiple- and single-pulse cases, the peak optical electric field incident onthe vesicles was 6-7x10 5 V/cm, about an order of magnitude above the membrarepotentials. -It appears likely that direct electrostrictive forces disrupted thevesicle membranes and facilitated their transition to stacked lamellar stPLL-tures known as liposomes..

Retinal damage thresholds in the Meteaca/fascicularis were determined forirradiation with single ultrashort pulses and entire pulse trains of 2nd harmon-ic light. In terms of irradiance at the retina, the fundoscopically deter-mined 24-hr postexposure thresholds were 4.4 mJ/cm 2 and 540 mJ/cm 2 respectively.The peak electric fields in both cases were of the order of 6-7x10 5 V/cm 2 , as inthe case of the vesicles. -Disruption of the cellular membranes is suggested sthe threshold damage mechanism. In addition, it is postulated that for pulsetrains, irreversible damage occurs very early in the train, at the first pulseto attain or exceed the threshold electric field.

The corneas of the same primates used in the retinal studies were irradiatedwith 4th harmonic (265 nm) mode-locked pulse trains derived from the Nd:Glasslaser. Fluorescein slit-lamp examinations at 24-hr post exposure revealed adamage threshold of 8.2 mJ/cm2 . All damage was confined to the corneal epi-theliun. It is postulated that photochemical processes such as coagulation ordenaturation of nucleoproteins and nucleic acids govern the damage mechanismin this case.

.. . . . . . . . . . .. . .. . . .. ... ... . . .. I . . .. ... ... . .. . . . . . . . III

PREFACE

The authors are deeply indebted to Ms. Carrol Harris for her expert

assistance with the primate eye irradiation studies. Ms. Harris carried out

all the veterinary procedures and assisted in the evaluation of the retinal

and corneal lesions. Thanks are also due to Mr. Nicholaus B. Martin for his

assistance in the vesicle studies and to Dr. Oktay Yesil and Mr. Russell Tom

for their technical assistance in the primate irradiation experiments.

Finally, the authors are grateful to Dr. John Taboada, U.S. Air Force

School of Aerospace Medicine, Laser Effects Branch, for carrying out the

probit analyses reported here.

\

1 i,

TABLE OF CONTENTS

Page

INTRODUCTION ............. .............................. 7

PART I: MACROMOLECULAR DAMAGE STUDIES ....... ................. 10

REVIEW OF PRIOR WORK ........... .......................... 10

OBJECTIVES ............. ............................... 11

SEARCH FOR PICOSECOND OPTICAL STRESS-INDUCED DAMAGE IN DNA ......... ... 12

Instrumentation and Techniques for Studying Possible Damageto DNA ........... ............................. .. 12

Gel Electrophoresis ....... ..................... ... 12Low-Shear Viscometry ....... ..................... ... 14Dynamic Light Scattering ...... ................... .. 15

Results of Damage Studies on DNA at 1060 nm ..... ............ 15

Damage Studies of DNA Irradiated at 1060 and 530 nmSimultaneously ......... ........................ ... 18

Summary of DNA Damage Studies ...... ................... .19

ATTEMPTS TO OBSERVE TRANSIENT DISTORTION BIREFRINGENCE OF DNA INDUCEDBY PICOSECOND PULSES OF 1060-nm LASER LIGHT ...... ............. 21

Theory of the Optical Kerr Effect in Liquids Comprised of SingleRigid Molecules ......... ......................... .. 21

Estimation of Molecular Parameters for CS2 . . . . . . . . . . . . 24

Theory of Induced Distortion Birefringence of DNA ........... ... 26

Attempts to Observe Transient Distortion Birefringence of DNA . 28

DAMAGE STUDIES OF LIPID BILAYER VESICLES ..... ................ ... 31

Dipalmitoyl Phosphatidyl Choline (DPPC) ...... .............. 31

Preparation of DPPC Vesicles ...... ................... .. 31

Experimental Measurements on DPPC Vesicles .... ............ ... 33

Results ........... .............................. .. 35

Egg-Yolk Lecithin (EYL) Vesicles ........ ................. 38

Preparation of EYL Vesicles ....... .................... .. 38

Damage Studies of EYL Vesicles ...... .................. .40

Summary of EYL Results ......... ...................... 44

Future Vesicle Work ......... ........................ .. 44

3

FPX1" PAGE BAMOI FL6

Page

PART II: PICOSECOND OCULAR DAMAGE STUDIES ON PRIMATES .. ......... .. 45

INTRODUCTION ............. ............................... 45

RETINAL DAMAGE THRESHOLDS INDUCED BY PICOSECOND 530-nm LIGHT PULSES 46

Irradiation Apparatus ........................ 46Configuration Used for Pulse Train Studies ... ........... .46Configuration Used for Single-Pulse Studies ... .......... .46Apparatus Common to Both Configurations ................ .49

Experimental Protocol ......... ........................ 50

Results ............. ............................... 53

Comparison with Other Work ....... ..................... .57

CORNEAL DAMAGE THRESHOLDS INDUCED BY PICOSECOND 265-nm LIGHT PULSES . . . 59

Irradiation Apparatus ......... ........................ 59

Experimental Protocol ......... ........................ 62

Results and Discussion ........ ....................... .62

REFERENCES ............ ................................ 66

APPENDIX A: PICOSECOND LASER IRRADIATION FACILITY ... ............ .71Nd:Glass Laser ....... ....................... .71Pockels Cell Pulse-Switching System ... ............ .74Pulse Chronometer System ....... ................. 75Pulse Energy Measurement ....... ................. 76

APPENDIX B: EXPERIMENTAL PROTOCOL FOR DNA STUDIES ... ............ .78

APPENDIX C: DYNAMIC LIGHT SCATTERING FACILITY .... .............. .. 79

List of Illustrations

Figure

I. Optical configuration for detection of picosecond birefringencefrom DNA base-pairs ........ ....................... .29

2. D vs. K2 plot for a sample with typical pulse-train irradiationof 600 mJ/cm2 total energy density ..... ................ .32

3. D vs. K2 for different preparations of EYL vesicles ............ 39

4. Schematic of apparatus for irradiation of primate eyes withentire trains of ultrashort 2nd harmonic (530 nm) pulsesderived from a mode-locked Nd:Glass laser . .... ........... 47

5. Schematic of apparatus for irradiation of primate eyes withsingle ultrashort 2nd harmoni-c (530 nm) pulses derived froma mode-locked Nd:Glass laser ...... ................... .48

6. Schematic of macular exposure sites ii, M. fascicularis retina .... 52

4

7. Schematic of apparatus for irradiation of primate eyes withultrashort 4th harmonic (265 nm) pulse trains derived froma mode-locked Nd:Glass laser ...... ................... .60

8. Schematic of exposure sites or M. fascicularis cornea ... ....... 63

A-I. Schematic of apparatus for picosecond laser irradiation ofmacromolecular systems ...... .. ...................... .72

A-2. Schematic of video detection and display system ............. .77

C--I. Experimental arrangement for dynamic light scattering ... ....... 80

C-2. Block diagram of digital clipped correlator .... ............ .82

List of Tables

Table

1. Summary of DNA irradiation experiments ..... ............... .16

2. Summary of DNA irradiation experiments at higher energy densities.. 17

3. Summary of two-wavelength DNA irradiation experiments(1060 nm + 530 nm) ......... ........................ 20

4. Summary of DPPC vesicle irradiation experiments ... .......... .33

5. Apparent diffusion coefficients (DxlO 8cm 2/sec) of DPPC vesiclesdetermined by dynamic light scattering .... .............. .36

6. Apparent diffusion coefficients of DPPC vesicles as functions ofsample age ............................................ 37

7. Values of D at various K2 for EYL vesicles .... ............. .412

8. Polydispersity P21T at !i=30' and u=120' for EYL vesicles ...... .43

9. Retinal damage thresholds at 530 nm ...... ................ 54

10. Comparison of threshold damage data for vesicles and retinas . . . . 56

]1. Comparison of present retinal thresholds with results of otherinvestigaticns (24 hr post exposure) ..... ............... .58

5

BIOLOGICAL DAMAGE THRESHOLD INDUCED BY ULTRASHORT FUNDAMENTAL, 2ND,

AND 4TH HARMONIC LIGHT PULSES FROM A MODE-LOCKED Nd:GLASS LASER

INTRODUCTION

The effects of laser radiation on various biological systems have been

studied extensively for several years (1-5). The eye has been found to be

particularly susceptible to damage by exposure to radiation ranging from the

ultraviolet (UV) to the infrared (IR) portions of the optical spectrum (1-3,

6-31). Depending on the wavelength of the laser radiation, ocular damage is

most likely to occur in the cornea, the lens, or the chorioretinal tissues of

the eye. The clear tissues of the eye are quite transparent in the visible

and near IR (32); consequently, radiation in this wavelength range affects

primarily the retina. In the UV the bulk of the incident radiation is absorbed

in the anterior portion of the eye; below about 300 nm, the absorption

occurs entirely in the cornea, usually within the corneal epithelium (2,25).

It follows then that ocular damage from the visible and near IR is sustained

primarily in the retinal tissues; while from the UV, it is sustained primarily

in the corneal tissues.

The mechanisms responsible for UV-induced damage in the cornea are

believed to be predominantly photochemical (25,27). However, relatively little

work has been done on the ocular hazards of UV laser radiation, particularly

in the ultrashort pulse regime and with respect to the effects of pulse dura-

tion in general.

By contrast, a large body of data exists on the effects of visible and

near-IR laser radiation on the retina (6-24, 28-31). The bulk of the work

has been carried out with CW or pulsed lasers of various wavelengths in the

nanosecond or longer pulse duration regime. Until recently, relatively little

attention was paid to the ocular hazards of ultrashort laser pulses (10-12 _10- 1 1 sec) (28-31). For the longer pulse durations (>10 - 8 sec), the injury to

the retinal tissue is believed to be caused by local temperature rise and

resulting protein denaturation and enzyme inactivation. Typical threshold

values of radiant exposure at the retina that cause observable lesions approach

7

EFECED1I PAi3R BL.&hC..oT mIIJ6

%I J/cm 2 as the exposure time is reduced to 10-8 sec. In the picosecond re-

gime, there is less agreement on the damage thresholds, and values as low as

2xlO -3 J/cm 2 (30) and as high as 2 J/cm 2 (28) have been quoted.

There is general agreement, however, that in the picosecond regime the

tissue damage is not of thermal origin (4,28-31). Several possible damage

mechanisms have been suggested, such as intense acoustic transients, direct

breakdown in the bulk, multiphoton ionization, and free radical formation

(4,24,28-31). Other possible mechanisms include direct electromagnetic field-

induced stresses in the nucleic acids, proteins, and cell and vesicle lipid

membranes (30,31,33,34).

Identification of specific damage mechanisms in living tissue is diffi-

cult because, unlike the optically pure samples used in inanimate material

studies, biological materials are heavily concentrated with optical inclusions,

free-charge regions, and dielectric absorption discontinuities. Furthermore,

the end point resulting from energy deposition may be reached by a complicated

set of pathological changes, with a result that is difficult to interpret.

The important constituents of a living cell that are likely candidates for

sites of radiation damage are its proteins, including fibrous structural and

contractile proteins, enzymes, and other polypeptides; its nucleic acids,

which are generally found in cells to be complexed with basic proteins (his-

tones) in a form called chromatin, or complexed with polyamines; and, finally,

its cell and vesicle lipid membranes (34).

We have approached this problem by (1) investigating the effects of ultra-

short-pulse laser radiation on isolated biological macromolecular systems

similar to constituents of living cells that are susceptible to damage and (2)

comparing the results to data obtained from direct ocular damage experiments.

The first phase of the program described in this report was designed to

study the effect of ultrashort pulses of 1060-nm light, and its 2nd harmonic,

derived from a mode-locked Nd:Glass laser, on aqueous suspensions of macromo-

lecular structures such as calf-thymus DNA, poly(L-lysine), DPPC vesicles, and

egg-yolk lecithin vesicles. The experiments have encompassed a wide range of

pulse energies for both single pulses and entire trains of mode-locked pulses.

Particle damage has been monitored by one or more of the following techniques

(depending on the particular macromolecule): dynamic light scattering, gel

electrophoresis, and low-shear viscometry. In the case of DNA, which suffered

L0

no damage detectable by these techniques at the energy densities attainable,

we have also carried out a thecretical and experimental investigation of

transient distortion birefringence to ascertain if any detectable internal

strain is induced by the laser pulses.

The second objective of our pogram was to determine direct ocular dan-

age thresholds in primates resulting from the 2nd harmonic (530 nm) and the

4th harmonic (265 nm) ultrashort-pulse radiation derived from the same laser,

and to compare the results with those of the macromolecular studies, with a

view to identify more closely the most likely damage mechanism(s). These

studies were also aimed to help establish new laser safety standards in the

picosecond time regime at these wavelengths.

This report is divided into two major sections: Part I deals with the

macromolecular irradiation studies, and Part II treats the experimental de-

termination of ocular damage thresholds in primate eyes.

9

PART I: MACROMOLECULAR DAMAGE STUDIES

REVIEW OF PRIOR WORK

Important macromolecular constituents of mammalian cells that are likely

candidates for initial sites of ultrashort-pulse laser radiation damage are:

(a) their proteins, including fibrous structural and contractile proteins,

enzymes, and other polypeptides; (b) their nucleic acids, which are general-

ly found in cells to be complexed with basic proteins (histones) in a form

called chromatin, or complexed with polyamines such as spermidine; and (c)

their noncovalent lipid membrane structures, including those of the outer

integument, various vesicles, mitochondria, and other endoplasmic inclusions.

Work described in our previous technical report on this subject (34) in-

dicated that the polypeptide poly(L-lysine) was unaffected by either single

ultrashort pulses of 1060-nm laser light with energies up to 1.37 mJ/cm 2 or

successions of up to four pulse trains, each consisting of -.100 pulses and2having integrated energy densities of up to 150 mJ/cm . It was inferred that

polypeptides, or proteins, are unlikely to be the initial sites of any bio-

logical damaae at these or lower energy densities.

The major effort of our previous investigations was to study the effects

of ultrashort pulses of 1060-nm laser radiation on DNA. Three potential dam-

age mechanisms were theoretically analyzed for their resulting distributions

of fragment sizes. We concluded that a clear-cut distinction between damage

mechanisms could in principle be made in the event that damage was unequivo-

cally observed. Unfortunately for us, but perhaps fortunately for laser-damage

victims, breakage of the DNA was unambiguously observed in only one sample, a

result that we now strongly suspect was spurious. Most of the experiments

revealed no detectable damage. Serious problems with both radiated and con-

trol YJA samole<> towar'd the end of the contract period resulted in a number

of awbiguous experiments in which tie DNA exhibited an anomalously high affin-

ity for the Millipore filters, with consequent slow filtration and loss of

sample and subsequent loss of precision in the dynamic light scattering

Pieasuret-it, which gave no positive lndic. tion of damaae in any case. Because

of thV'c, tec', n cd1 difficulties an, tecause -f hf importance of observing

ci rec:l 1 tflo fragment distribution, vw undertook the development of a newanalysis technique, gel-electroi)horesis. However, final completion of the

necessary apparatus and application of that technique to study irradiated DNA

samples had to await the present contract period.

In summary, no detectable damage of DNA was produced by single pulses up

to 0.75 mJ'cr", or two pulse trains up to 230 md/cm total energy density, or2

four pulse trains up to 440 mJ/cm . At higher total energy densities, theresults were ambiguous due to technical problems. A primary objective of the

present work has been to obtain unambiguous results for DNA irradiated at the

highest obtainable total energy densities.

Some initial work was performed on the physical characterization of a

preparation of purple membranes from Halobacterium halobium, which we hoped

would provide a tractable model for a retinal membrane-protein system. Un-

fortunately, the ultrasonically dispersed suspensions of this material were

neither as reproducible nor as stable as desired, so no irradiation experi-

ments were undertaken.

OBJECTIVES

At the outset of the present investigations we intended to study DNA ir-

radiated at the highest attainable energy densities--using dynamic light

scattering, low-shear viscometry, and especially, gel electrophoresis--to

detect any damage sustained and to analyze the distribution of product frag-

ments. When it became clear that DNA was not detectably damaged byeven the

most energetic available laser pulses or whole trains of such pulses, we de-

cided to investigate the transient distortion birefringence induced by such

picosecond pulses. Our idea was that, even though rupture of covalent bonds

was not observed, we might still be able to observe transient distortion in-

duced by the high-power optical pulse; however, no transient distortion bire-

fringence was ever observed.

When it became clear that the covalent bonds of both polypeptides and

nucleic acids were unaffected by ultrashort pulses of 1060-nm light at the

energy densities attained in these studies, we decided to focus our effort on

noncovalent lipid membrane vesicles. Indeed, it was unambiguously and repro-

ducibly observed that irreversible changes were induced in suspensions of di-

palmitoyl phosphatidyl choline vesicles by trains of pulses with total eneroy2densities exceeding 600 mJ/cm

11

SEARCH FOR PICOSECOND OPTICAL STRESS-INDUCED DAMAGE IN DNA

The quality control problems with the calf thymus DNA and/or filtration

system that surfaced in the final phase of the early work have been essenti-

ally overcome by a combination of procedures, including

(1) washing the filters in a hot solution containing 0.0115 M Na2CO3 and

0.0015 EDTA, and

(2) working with the DNA in a higher pH buffer, I M NaCl, 0.01 M EDTA,

and 0.05 M Na2CO3, pH 9.3.

We do not understand why DNA passes through Millipore filters so much more

easily at high pH (> 8.5) than at neutral pH, but that has been a repeated

observation in this laboratory.

Laser exoeriments with DNA at 1060 nm were performed following the pro-

tocol described in our previous report (34) with a somewhat modified laser

system. Descriptions of the irradiation apparatus and experimental protocol

are reproduced here in Appendixes A and B respectively. We temporarily re-

moved the optics used for pulse chronography in order to increase the pulse

energy available at the irradiation sample cell. The laser was operated

multimode with the larger diameter laser beam focused down with a lens system

to match the ID of the sample cell. This further increased the irradiation

intensity. The beam profile, although not Gaussian, was free of large-scale

intensity fluctuations (i.e., hot spots). Spatial intensity ripples were

within 10% of the mean local intensity.

Instrumentation and Techniques for Studying Possible Damage to DNA

Gel Electrophoresis--A gel electrophoresis apparatus was constructed

using a J-wick flat-bed configuration designed for very dilute (0.5 K or less)

agarose gels. The entire unit was housed in a Lucite safety box with micro-

,witches in the latch to guarantee interruption of the hich-voltaqe leads

whenever the box was open. Considerable experimentation was done with methods

of preparing, pouring, storing, loading, running, and developing the dilute

agaros gels. Some of the most important guidelines to emerge from this

activity were the following:

(1) Agarose concentrations of at least 0.3% (w/V) were required to formgels strong enough to permit the necessary manipulation withoutbreaking.

(2) A "frame" several mm wide and 1-2 mm deep of more concentrated gel(1%) on the periphery was essential to provide adequate mechanicalstability for manipulation of the more dilute bulk gel.

(3) Interchanging the buffer in the anode and cathode compartmentsevery few hours was effective in moderating pH drift.

(4) Forming the gel on top of a removable plastic slab greatly facili-tated the subsequent manipulations.

Typically, 10-20 microliters of solution containing 0.03-0.10 mg/ml DNA

were loaded into one of the wells, or slots, on the "starting line" of a cold

gel. These wells were impressions of the embedded teeth of a suspended comb

remaining after tie gel had cooled. The loaded wells were capped with a drop

of warm agarose solution that was allowed to harden. The electrophoresis

buffer that we have found works best, and gives the least pH drift, is 40 mm

Tris (base), I mM NaEDTA, and 5 mM H3BO3. The voltage was initially set at

60 kV for 20-30 min until the DNA penetrated into the gel (from the well),

then was maintained at 30 kV thereafter, which produced about 2 mA current.

Bromthymol blue was added to one of the wells as a migration indicator. It

moves with two to three times the speed of the native intact DNA. Typical

run times were 18-24 hr. The gel slab was then removed from the apparatus and

soaked in a solution containing ethidium bromide (%0.001 mg/ml) for 20-30 min.

Then the gel was transferred to the illuminating upper surface of a UV black-

light box. Ethidium cation fluoresces with a quantum efficiency about 200 x

higher when bound to DNA than when free in solution, thus its bright orange

fluorescence locates the position of the DNA "bands" in the gel. These bands

could be photographed using an appropriate filter to block the exciting UV

light. The resulting negatives could be scanned by densitometer to provide

a very precise quantitative picture of the DNA distributions in the bands.

Although this was done for a few runs, using equipment in the laboratory of

W. Fangman, direct visual inspection generally sufficed to demonstrate that

no damage was occurring. Therefore, densitometry was not further pursued,

and that subject will not be discussed in the sequel.

At the time of writing, the total number of DNA samples analyzed by gel

electrophoresis in this laboratory (for this and other projects) is well over

13

100. We have found that the technique is extraordinarily sensitive for de-

tecting double-strand breaks, and practically, if not totally, insensitive to

sinQle-strand breaks. A surprising finding of our ongoing investigations of

DNA was that even the internal motions manifested in the dynamic light scat-

tering were completely insensitive to single-strand breaks at neutral pH (35).

In fact, data in the older literature (36) indicate that the sedimentation

coefficient and radius of gyration obtained from static light scattering, both

of which monitor the bending rigidity, are insensitive to single-strand breaks

at neutral pH. Thus, the insensitivity of the gel electrophoresis to single-

strand breaks is consistent with these other observations. A recent discovery

has shown that single-strand breaks, bound protein contaminants, and bound

spermidine, can all induce profound changes in the Rouse-Zimm model internal

motion parameters at pH 10.2 (37). These facts may be useful in any future

DNA damage studies.

The presence of double-strand breaks, which we have often observed in

samples contaminated with nuclease activity, is manifested by lower molecular

weight components migrating, or streaking well ahead of the main band of the

intact DNA. A simple description of the gel electrophoresis technique as

applied to high molecular weight DNAs is given by Fangman (38).

Low-Shear Viscometry--A Crothers-Zimm-type rotating float viscometer

with a shear of 0.56 sec - 1 has been used to analyze irradiated DNA. The low-

shear specific viscosity, as defined by

nsolution

nsolvent

is approximately proportional to M where M = molecular weight, for

a fixed weight concentration of DNA, and thus provides a rather sensitive

indication of the average molecular weight in the sample (36). Considerable

experimentation was required to evolve a procedure that would yield repro-

ducible and reliable values of the accuracy required. Nutation of the float

greatly perturbed the weasured viscosities, and eventually was controlled

only by altering the configuration of the rotating magnets, careful monitor-

i. of the sample volumTle, and scrupulous attention to cleanliness.

The intrinsic viscosity FT.! = /c = 6) dl/q observed for the control

is close to that expected for DNA molecules, of molecular weight 12x10 6 , which

is whit good samoles of calf t~ywus DNA nearly always turn out to be.

Dynanjic Lih5catterinj --The experimental instrumentation and tech-

niques, including filtration Procedures, were exactly as described in our re-

cent report (34). (See also Appendix C.) This was by far the most difficult

and time-consuming assay for damage, and was often fraught with problems be-

lieved to arise from dust inevitably present in the sample irradiation cells,

which were very difficult to clean. Thererore, this technique assumed a

secondary importance to the gel electrophoresis technique and was not run on

all samples, especially those exhibiting significant dust problems.

Results of Damage Studies on DNA at 1060 nm

Table 1 presents the results obtained for five separate samples of the

same calf-thymus DNA, which were exposed to picosecond pulse-envelopes ranging

from a single pulse up to four pulse-trains, each containing the number of

pulses indicated. The maximum energy densities are significantly higher thanthose reported previously. The differences in properties of the irradiated

and control samples lie well within their respective experimental errors in

all cases. The positions and widths of the gel electrophoresis bands of all

the irradiated samples were identical to those of the control, as denoted by

the "neg" result. There is clearly no significant fragmentation of DNA at

these power levels under the present conditions.

The presence of unusually large amounts of dust and/or aggregates in

these samples precluded accurate dynamic light scattering measurements, but

the gel electrophoresis was fortunately unaffected by such contaminants. In-

terestingly, the occurrence of Tyndalls due to dust and/or aggregates dimin-

ished in the order in which the samples were introduced to the cell, suggest-

ingstrongly that the dust was initially present in the radiation cell, which

was very difficult to adequately clean, and was gradually washed out by suc-

cessive samples. The control, which was run last, was quite clean.

Table 2 presents results obtained for calf-thymus DNA at still higher

energy densities. In fact, in sample 5 of that Table the incident beam was

focused into the sample cell with a 55-mm-focal-length (f.l.) lens to cause

dielectric breakdown in the DNA solution. The focal spot size in that case

was -55 om. The sample was irradiated with two pulse trains and was reposi-

tioned after the first train so that breakdown would be produced at a differ-

ent place in the cell. The occurrence of breakdown was evidenced by the

generation of a greenish spark ,2 mm long and a sharp audible acoustic

15

0 0 000

0.-

(CL;)I ta (1 ) a) ) 0

C Dm) Lo C\J

U) a)" I I

C 4- (\j~) E

V) - 0 0

CL 00 \J LA0

'4J -a I I IIM: r- r-

CCD

L u)4-4 .4-)

0 (a,

;2f LO t.0

LL- CU)C

C) L)

4-))

>4 t>

CU E LU

co C\J 0.'oC: O r--d CjW EL ~ lL 1 -.-- r. - -1- LO U)j

U L

LU ~ ~ ~ ~ V >., -' *U ' .Ck) n~

C0) ECL in

4/) 4/) C/) 0)

115 7 l s- S.. - -fl 4-

0- 44.C

Lo iL LUJ

-~ \) ') IT Ln

V)'

(o

S.-

L/) 0LIII S -4 C) 000

LL] 4-) a WCD ~ ~u -1

-(A C\i

*-C)

(I) -) 00 -t

UL >(O

x. CD CD C -4 -4-

->1 . XX

C)'

C)CL

V) 0, roC4- L-4 a) ) LM

V) a. -4-Nd4

(.0 )<. m

-40 o S S.:- (D- -

S.- 4- a - S

0- 0

.*- (A *.- -(A C -

17

transient. Maximum electric-field strengths in these focused pulse trains were

about 6 x 109 V/m. Again, both dynamic light scattering and gel electrophoresis

indicate that there was no significant fragmentation of any of these irradi-

ated DNA samples. In related noncontract research, we have found that the

apparent translational diffusion coefficient D(0=250) is much more sensitive

to DNA fragmentation by contaminating nucleases than is D(0=90').

It may be concluded that there is no detectable fragmentation of DNA upon

irradiation by single picosecond pulses of 1060-nm laser light with energy2

densities up to 17 mJ/cm , or pulse trains with energy densities up to22.16 J/cm , or focused pulse trains with energy densities at focus up to

4.5 x 103 J/cm2 , the latter being sufficient to produce local dielectric

breakdown.

Damage Studies of DNA Irradiated at 1060 and 530 nm Simultaneously

Letokhov (39) has suggested that multistage excitation of DNA in solution

to energy levels high enough to produce dissociation, or fragmentation, could

be achieved by simultaneous irradiation with very intense ("109 W/cm 2) pico-

second (',10-ll - 10-12 sec) pulses of one or more frequencies, either one of

which is too low to photodissociate the molecule in a single quantum event.

This viewpoint ignores the effects of anharmonicity in the vibrational mani-

fold and assumes a 10-fold smaller relaxation rate (at high excitation levels)

than the well-documented very rapid relaxation of energy among the numerous

(coupled) nondissociative modes at high excitation levels of large molecules

even in the gas phase (40). The availability of a great many more coupled

modes to accept the energy in condensed phases will probably increase the

power requirements for such multistep photofragmentation far above the value

estimated by Letokhov. Nonetheless, we decided to investigate the effect of

irradiation at 1060 nm and 530 nm simultaneously. Two 530-nm photons would

be equivalent to one at 265 nm, close to the maximum of the DNA ultraviolet

absorption band at 263 nm.

The experimental setup was similar to that used for the 1060-nm studies,

except that the dichroic beamsplitter tiat normally separates the IR and

green components was replaced with an uncoated-glass beamsplitter. Thus, the

2nd harmonic could reach the target cell. Relative measurements of IR and

green pulse energy for two experiments yielded an SHG conversion efficiency

of .3' in one case and 17 in the other. The D.NA samples were contained in

tht. 5-rv, - iD x ?-r-mw quartz cells. The la,.cer c:,erture was set at 8 mm and a

50-cm-f.l. lens was used to reduce the beam diameter to 5 mm. By measuring

burn spots on Polaroid film at the location of the target cell, we determined

that the diameter of the 2nd harmonic beam was about half that of the IR.

This is due partly to the square-law harmonic generation and the different

effective focal length of the focusing lens for the two different wave-

lengths.

Table 3 contains the results of these simultaneous (1060 nm + 530 nm)

irradiation experiments. Again, it may be concluded that no fragmentationof the DNA was detected. Although sample 1 seems to be giving slightly high-er apparent diffusion coefficients, its gel electrophoresis band was unequiv-

ocally negative. As was often the case, the presence of dust in these

samples greatly reduced the accuracy and reliability of the dynamic light

scattering results, especially aL the lower angles.

Summary of DNA Damage Studies

Tables 1, 2, and 3 of this report and the Summary Table of our previous

report (34) show that our data overwhelmingly indicate no detectable damage

in irradiated DNAs under any conditions that we have studied. The sole ex-

periment that indicated positive breakage has now been superceded by numerous

experiments at much higher energy densities, all of which give unequivocal

negative results. Therefore, that single result is now presumed to be

spurious.

Numerous direct attempts to produce a damaged DNA similar to that ob-

served in the spurious experiment by conceivable experimental errors, such

as deliberately leaving residual chromic acid cleaning solution in the radi-

ation cell, were unsuccessful. However, in the course of subsequent noncon-

tract research, we have actually twice experienced similar damage in DNAs

stored in dilute solution in our refrigerator, presumably a consequence of

the rapid production of nucleases by contaminating cold-insensitive micro-

organisms. We are, therefore, strongly inclined to ignore that one experi-

ment, and accept the obvious conclusion that irradiated DNAs suffer no dam-

age detectable by our techniques at the energy densities attainable.

19

LiZ 0 0

aI)S.-

0 (1

H Ci

4-) o 0

4- U

CL Ej a)J(0 LA0m

oL k

44--

00

cx C~ I~

0~~. 0. Ic ) \I- LfO LO-j ~LO

-LJ C2

<i C

C)CC: -C4 ) \JC

LL- Q U-EC)'

0) E5.-C

Li M (n 0V.) LO U n

2i E

co C C C

(V E

S..-

4-'I

0 0

ATTEMPTS TO OBSERVE TRANSIENT DISTORTION BIREFRINGENCE OF DNAINDUCED BY PICOSECOND PULSES OF 1060-nm LASER LIGHT

One of the potential damage mechanisms for DNA conjectured in the previ-

Gus report (34) was based on intramolecular strain caused by light-induced

torques on the optically anisotropic DNA bases. Even though strain to the

point of fragmentation did not occur, strain may still take place to a suffi-

cient extent to produce a detectable birefringence. An experiment to test

this hypothesis is worthwhile for several reasons. If observed, such an in-tramolecular strain might permit a rough estimate of the laser powers ulti-

riately required for fragmentation. Also, if observed, the transient decay of

the optically induced birefringence could provide new information regarding

the dynamics of these macromolecules on the picosecond time scale. Even ifnot observed, it might be possible to set a useful lower limit on the re-

straining forces governing tilting of the DNA bases with respect to the helix

axis, about which essentially nothing is known.

Theory of the Optical Kerr Effect in Liquids Comprised

of Single Rigid Molecules

In a liquid comprised of cylindrically symmetric nonpolar molecules with

oolarizability ao = ('X11 + 2cx,)/3 and anisotropy Aa = l- L' the dipolemoment (.zi) of the ith molecule induced by a z-polarized constant macroscopic

electric field (E z) can be written in the form

z= fEz[uoPo(cos 'i) + 2 A0 P2 (cos 0i)], (1)

where r. is the polar angle between the molecular symmetry axis and the elect-icfield, P..(cos *.i) is the ;th Legendre polynomial of cos -i' and f = Eloc/E z is

the internal field correction factor relating the local cavity field Eloc to

EzInteraction energy between the applied electric field Ez 0 rEz and the

total induced polarization in the sample of N molecules in volume V is (41)

0 (Ez°) 2H(t) Ez dpz . ... 2 f[N't° + - An j P2 (cos 0j)i. (2)

2 1

21

Neglecting the orientation independent part, and generalizing to the mean in-

teraction energy with a very rapidly oscillating optical field (for which

(Ez°)2/2 -1 (E o)2/4), this relation takes the canonical form of Kubo (42):z z

H(t) = -F(t) • A = -Act)' P (Cos 6 (3)4 3

where the second equality serves to define the quantities F(t) and A. The

time-dependence of F(t) - (Ez 0)2/4 is assumed to be very slow compared to the

optical frequency of the irradiating light.

Using the relations Dz Ez + 4vP z = czEz , n z (at optical frequen-

cies), and nz = n0 + In, it is readily found to lowest order in Aa that

nz = 0 L- ZV P2 (cos e ) = B, (4)

where the last equality serves to define B : 6nz.

From Kubo's fluctuation-dissipation theory (42), the linear response of

B = .nz to a sinusoidally varying F(t) = [(Ez°)2/4]cos wt (where u, is here

understood to be much smaller than optical frequencies) is simply

B(t) = Re.: BA("') Fo cos £,t + Im{,BA()F o sin (t , (5)

where the real and imaginary parts of the susceptibility are given by

Re{yBA( "k= kBT AB(O)> W k AB(t)> sin fut dt (6)

0

Im1 BA( , kBT 4AB(t), cos . dt. (7)'B T

Here T is the absolute temperature, kB is Boltzmann's constant, and the angu-

lar brackets denote equilibrium averages for the system in the absence of the

perturbinq optical field.

22

For F(t) varying slowly compared to the rotational relaxation time (e.g.,

of CS2 in the present instance), Iv, BA(.') } = 0, and Re fB(")} B AB>/kBT.

Thus the "steady state" (i.e., .=0) response of the refractive index to the

perturbing optical pulse is2

(E ) f2 8 - 1

4 kB 9 P2(cos i P2(cos 9). (8)0! ij

Now, using the identity P2 (cos i Y2 0('i), where Qi = (.,Oi), and

Y m *) is the corresponding spherical harmonic, one obtains.m4 < (2j) P U uj)>, (9)

<P2 (cos ) P2 (cos 5).= -- <Y20(oi)Y 20 = 5 u2i , (

where the last line follows from the addition theorem of spherical harmonics

(43) and the fact that the liquid exhibits an isotropic equilibrium state.

The <i' ;j are unit vectors directed along the symmetry axes of molecules i

and j.

The instantaneous angular distribution of symmetry axes can always be

expanded in the complete set of Legendre polynomials. It is apparent from

Eq. 4 that 6nz must be proportional to the coefficient c2(t) of the P2 (cos

term, provided that c2(t) vanishes in the equilibrium state and does not con-

tribute to n0 , as in the case for i! otropic fluids. Under the same conditions,

still to lowest order in .. ,, it is readily shown that 6nx =ny = -2 nz.

Thus, one obtains finally_n-fn -~ = 3An z

An = 6nz nx z

(10)2 (u) 2 2

Sn k BT 1 {5 [f2 {l + (N-l) < P2( i • C2)>}] (Ez°)

where (E z)2 is the square of the (peak) amplitude of the perturbing optical

beam in vacuo. This formula differs from the standard formula applicable to

dilute gas phase molecules,

23

2V Ag A,x C_ nAn = 2 (Ez)2

15 n2kBT Z1no kBT

2 : (_ V ) A (I (01 )

15no k BT (Ez

by the appearance of the factor in square brackets. In Eq. 11, C is the3 gconcentration in g/cm , Ag = Aa/Vm, where Vm = molecular volume in cm , and

V = specific volume (cm per g); hence C9 (V/Vm) = N/V.

The Clausius-Mosotti theory of the internal field gives

1_ 3 V 1 no0 3 (. M ) noL

0o N 2 4T Av i + 2 (12)0 o

Jicre M, NAv, and p are the molecular weight in daltons, Avogardro's number,

and the density, respectively. The same internal field theory, which is

known to be only approximate, gives

fEloc _no 2 + 2

E z n 3 (13)

Unfortunately, the value of the orientation pair correlation function

OPC2 F fI + (N - I)<P2(i i2)>} is not available for CS2, and no simple

theory is available for guidance. Values reported for chloroform and nitro-

benzene are 2.0 and 2.8, respectively.

Estimation of Molecular Parameters for CS2

For CS2 at 25%', the Handbook of Chemistry and Physics gives a refractive

index n = 1.628 and a density p = 1.595 g/cm 3 . Using these values in Eq. 12

above gives ao = 6.7 x 0 cm. The value f = 1.55 is estimated from Eq. 13.

The birefringence induced by high-power optical pulses has been found

experimentally to obey the relation (44)

24

zAn = 2.2 x io-2 0 (Vol t/m)M

= 2.0 x 10-11 (statvol 2= Its/cl")RMS

= 1.0 x 10-11 (Ez 0)2 (14)

Comparing this expression with Eq. 10 and using the values of n0, a' f, and

N/V = PNAV/M given above, yields finally

7.25 x 10-2 4 3:1I PC cm. (15)(OPC2)

The interesting difficulty here is that if OPC2 were of order 1.0, then

t- would be comparable to, or even larger than, the estimated o = 6.7 x

10 24 cm3 . A second estimate of q is available from the molar polarization3o

(4u/3)NAv'o = 21.1 cm 3/mole of CS2 dissolved in other nonpolar solvents, as

tabulated by Debye (45). This datum gives . = 8.4 x 10 2 4cm- , comparable

to the previous value. One is forced to conclude that either Aa or OPC2, or

both, are much larger than expected. Typically Aa - aII -L is only a few

percent of r 0 (1/3)(,11 + 2ct ). However, for CS2, we must have Ac/cx0greater than 0.3 if OPC2 is less than 10!

The point of the above discussion is to note that CS2 provides an extra-

ordinarily large induced birefringence due to either an enormously large Aa

in comparison to or to a very large equilibrium orientation pair correla-

tion function, which implies that many CS2 molecules will orient cooperatively

due to strong intrinsic orientational coupling among them, or to both effects.

From a comparison of Eqs. 10 and 11, we see that the simple standard formula

in Eq. 11 can be used if CS2 is assigned an apparent ,a app defined by

, (app '41 55) (OPC2)

1.2 X 10-23 C13 (16)

25

where use has been made of Eq. 15. This apparent anisotropy substantially

exceeds q for the reasons cited above.

Theory of Induced Distortion Birefringence of DNA

An optical pulse of duration 10- 1 1 sec passes by too rapidly to effect

significant reorientation of the helix-axis of DNA but may be able to tilt

the base-pairs (which are regarded here asasingle rigid unit) with respect

to the helix-axis, thereby inducing some distortion birefringence.

Attention is focused on a single base-pair, which is imagined to have an

optical polarizability that is cylindrically symmetric about the fixed local

helix-axis in the undistorted configuration. That fixed helix-axis in turn

makes an angle e with respect to the polarization vector z of the optical

pulse. If there were no distortion, the interaction energy of the single

base-pair with the optical field would be given by the appropriate modifica-

tion of Eq. 2 for a single molecule.

The orientation of the base-pair with respect to the helix-axis is gov-

erned by an angular harmonic restoring potential gr 2 /2, where g is the restor-

ing torque constant and n is the angle of deflection or tilt away from the

eQuilibrium configuration. The equilibrium probability distribution for n is

2_gn

P(n)dn = e 2kBT dr. (17)/2 BT/g

The interaction energy between the optical pulse and the base-pair is

o)(E 0)2H(t) = - {- o 2 P P2 [cos ( - n)]} (18)

If g is sufficiently large, then r - 1. Direct expansion for small r, < I

gives

2P2[cos(o - n)] (1-2Tj )P2[cos o] + -- + 3 r cos sin q (19)21s( 2 2

26

Because the helix-axes are held fixed, the system can respond to the optical

pulse only by changing -. Thus, the interaction energy between the optical

pulse and those degrees of freedom (i.e., rl) that are free to respond is

F(E 0)2 2 2S(t) = - .j L 4 f { -2, 2 [cos )] + r, + 3 T. COS sin i

= -F(t).A . (20)

The change in refractive index induced by the pulse acting on this one

base-pair is found from Eq. 4 in the form

4T- f

6nz 3 n V P2 [cos (0 - )]. (21)0

Because the a value for a given base-pair is fixed, only that part of n

that varies with n will actually contribute to a 6n induced by base-tilting.

Therefore we keep only the lowest rn-dependent terms

6 n f -- t-2n2 P2 (cos :.) + + 3n sin 0 cos e}'S 3 n 0V2

SB (22)n

Again, linear response theory may be employed with the result that the

steady-state birefringence is given by

F(E Zo) 2-<AB >nz L 4 kBT

(23)

(E 0) f 2 87T A.2 22 2 2E 2 8 2 P2[cos (0] + a- + 3n sin 0 cos j2>4 k BT 9 - V 2 2

where the angular brackets denote an average over n using the canonical dis-

tribution in Eq. 17. Odd powers of r, average to zero, and fourth-order terms2are neglected; so only the = kBT/g 1 terms contribute. It is assumed

that we are in the stiff limit where g *> kBT. Consideration of a collection

27

of N independently tilting base-pairs can be accommodated by multiplying Eq.

23 by N and performing an angular average over the uniform distribution of

helix-c is orientations o. The entire final contribution comes from the

<(3r, sin ,cos 1)2> term, and is giver, by

A 4i" f2 Au zN)2on - A () (E). (24)z 15 _g Pn Ez0)(4

Again, An = 1-- n. 3,, so finally

L2

n 4:f 2 A 2 () (Ez0) (25)5 g n V z

which may be compared to Eqs. 10 and 11 for freely orienting species. It

should be noted that Eqs. 24 and 25 apply only when g/kBT -.- 1.0.

Attempts to Observe Transient Distortion Birefringence of DNA

Herring sperm DNA (Na salt), Type VII from Sigma Lot #67C-00251, was

sheared and concentrated down to approximately 63 mg/ml using a french press

at 12,000 psi, and then evaporating in a vacuum. The size of the sheared

fragments was estimated to be roughly 200,000 daltons with an average diffu-

sion coefficient D z 8.2 x 10-8 cm2/s. Shearing was necessary to achieve such

a highly concentrated solution. This was then used as the active medium in an

optical Kerr effect streak shutter (46), as indicated in Figure 1. Unfortu-

nately, no detectable birefrinqence of the DNA was ever observed with DNA as

the active medium in this apparatus.

The intensity of the maximum transmitted signal appeared as a vertical

deflection of 5 cm on the oscilloscope display. A deflection as small as 2 mm

would have been readily observed. In addition, the probe beam intensity was

increased 60-fold for the DNA experiment. Thus, the transmission sensitivity

of the streak shutter was Gnx50/2 = 1500 times higlher in the DNA experiment

than for CS2. Because the particular apparatus used in this streak shutter2

gives a transmitted intensity signal proportional to /n , we may conclude

that

28

UlU,

CD-

C))

UU4A

0

LL..

0 U

CD m

4-

o CL

CDC

Q)>- Lai >-

ui0

29U

Amax 1 Anmax n max (DNA CS 38.7 CSL)

25 22

From the measured intensity of the gating pulse (I,109 W/cm2), the RMS volts/rn. ax -

can be recalculated and used in Eq. 14 to give mA nc 8xi0 5 . Thus,

An max ' - S2DNA - 2x1O 6

A lower limit on the torque constant g can be obtained from Eq. 26 bynma x max

using Eqs. 11 and 16 for on the right-hand side, and Eq. 24 for AnDNA on

the left. That is,

4r f2 ('l NAv)E o2 1 2Tr 1 2 PNAv 0)2

38.7) Ao. M ~) (E 0) (27)nogh0 M z - 38.7 I-5 kBT app--MEz2

where the tilde (-) denotes quantities pertinent to the DNA solution. The

oarameters for CS2 were given earlier. For water, n 0 1.334, and Eq. 13

gives f = 1.26. For the DNA solution, = 0.063 a/l, and M = 662 per base-

pair. For one base-pair we take Ai = -14.1x10-24cn, which we have calculated

from the flow birefringence data of Harrington (47) in 0.2 M NaCl, using his0

Eq. 2 together with the persistence a = 490Areported by Voorduow et al. (4F).

This value is slightly higher than that (Aa = -12.9xl- 24 crr) computed byHarrington using a = 660 A. Using these values, we obtain,finally, the lower

limit

g (3.23) kBT = 1.32xi0 "13 dyne-cm. (28)

This lower limit only marginally satisfies the inequality g >> kBT required

for validity of Eqs. 20-25, and in that sense is disappointingly small. Thetorsion constant between base-pairs for the twisting of DNA has recently been

measured (49) and found to be 3.8x10 12 dyne-cm. Thus, this experiment has

established only that the torque constant for base-tilting must exceed approx-

imately 1/30 of the torsion constant for twisting successive base-pairs

(about the the helix axis) with respect to one another. That is neither a

surprising nor a very useful result.

The sime DNA preparation likewise showed no induced birefringence after

being immersed in boiling water for 20 min (to denature the DNA) and then

quenrched in an ice bath.

30J

We note that i the intensity of the orientini ct C1 eilse could be

increased by a factor of 10, then even a negative refult would provide an

interesting lower limit for the base-tilting torque constant.

DAMAGE STUDIES OF LIPiD BLAYER VES1CLES

Dipalmitoyl Phosphatidyl Choline (DPPC)

This synthetic pure lipid (i.e., DPPC) forms bilayer vesicle membrane

structures by appropriate preparation, or introduction, in aqueous solution.

It does exhibit a thermal order-disorder transition in the bilayer near

Tm = 41'C. The experiments reported here refer to vesicles at T < Tm, which

exhibit the more ordered, less fluid phase of the membrane bilayer. They are,

therefore, probably less representative of biological membranes, which are

composed of a variety of mixed lipids for which the fluid phase is believed

to prevail at room temperature. A few experiments performed subsequent to the

contract indicate that the stabi ity of DPPC vesicle suspensions is consider-

ably reduced for T _ Tm

Preparation of DPPC Vesicles

Vesicles were prepared from a 95, ethanol solution of dipalmitoyl phos-

phatidyl choline (DPPC) a. a concentration of 23 limole/ml (50). A 0.50-ml

aliquot of the ethanolic lipid solution was injected into 10 ml of buffer

(0.1 m NaCl and 0.01 M Tris-HCl) at a rate of 0.20 ml/min to give a final

lipid concentration of 2.3 pmole/ml. This injection method produced vesicles

about 400 A in radius, in good agreement with literature values (51). Some

polydispersity existed even before irradiation, as evidenced by a decline in

the apparent diffusion constant at low scattering angles (e f 45-1, or

K2 < 1.0 x 1010 cm-2 ), as shown in Figure 2. After exposure to picosecond

laser radiation, the 2.3 pmole/ml sample solution was transferred directly in-

to a light scattering cell that had been previously rinsed free of dust by

washing with filtered (0.2 um Millipore) buffer. The sample solution was

diluted with filtered (0.2 wm) buffer to 0.23 omole/ml for use in dynamic

liniht scattering (DLS).

31

LUi

U) E

<Li0 4-0

<C .- ~ c

ooUZ( 4-)

CEECE0 >0

X~~~ E - 4'0

.- a) C

Eo a

020S-0

KI 04- '

4-'

C: S.-

n

-.>

00 CC

o6C LO

cu

o)s 0o 0 0

Experimental Measurements on DPPC Vesicles

Nine successful exposures of the vesicles were carried out ranging from

an energy density of 9 mj/cm 2 in a single pulse to 3 J/cm 2 in an entire pulse

train. The exposure parameters of the various samples are given in Table 4.

TABLE 4. SUMMARY OF DPPC VESICLE IRRADIATION EXPERIMENTS

Energydensity

Sample Irradiation (mJ/cm ) Damage

1 pulse train 706 yes(massive)

2 1 pulse 9.0 possibly(slight)

3 pulse train 580 yes4 pulse train 614 no5 pulse train 3,080 yes

(slight)6 pulse train 529 no7 pulse train 533 ambiguous

(probably no)8 pulse train 533 no9 pulse train 529 no

Studies with DLS were performed on all of the samples as well as their

appropriate controls. For a polydisperse solution of scatterers with dimen-

sions small compared to K-I, where K - (4,n/.)sin 0/2 is the scattering vec-

tor, n the index of refraction, and o the scattering angle. The apparent dif-

fusion coefficient obtained from the dynamic light scattering is the so-

called z-average value (52) defined by

0m2 2 (29)

where ni , mi, and Di are the number concentration, mass, and diffusion coeffi-

cient of the ith species. For spherical particles of dimension comparable to

or larger than K-. the appropriate generalization of Eq. 29 is

iapp nI2 i (K)D/ n j (K), (30)

33

where Pi(K) is the internal interference factor characterizing the ith species.Pi(K) decreases from a maximum value of 1.0 at small K2 to a lower value (de-

pending on the particular shape and size of the scatterer) at any given larger

value of K2. For sufficiently large particles, Pi(K) is negligible at largern~ 2

K , but at small K may permit the large particles to dominate the scattering

from the smaller species, thus causing a decrease in the apparent diffusion

coefficient at small K2 (as indicated in Fig. 2). Conversely, a significant

decline in D at small K2 implies the existence of very large scatterers in the

preparation.

The apparent Stokes radius Rh = kBT/67nD--where kB is Boltzmann's constant,

T the absolute temperature, n the solution viscosity, and D the diffusion coef-

ficient--was computed for the very large particles from the low-angle (250 or

30') data for comparison with static light scattering measurements of the radii

of gyration R of those same particles. The latter values were computed usingg 22

the relation applicable for k R2 a 1.0

IN(I)/INO) : 1 - K22/3, (31)

where IN(O) is the "normalized" intensity (i.e., corrected for the relative

size of scattering volume subtended at n), and IN(0) is the extrapolated

value of I N(o) at i = 0. A comparison of Rh and R values obtained for either

radiated samples or controls over the mid- to low-angle range of measurements

shows that Rg > Rh (typically, R, Z 1,600 A, Rh z 720 A), which implies a

nonspherical, irregular or anisometric shape for particles of very large di-

mension, thus indicating that they are not simply very large spherical

vesicles.

The ordinary vesicles (Rh z 450 A) do not satisfy the criterion,

K2R2/3 1, and hence do not produce sufficient variation in IN() with to

yield a meaningful estimation of R using the present data. In experimentsgsubsequent to the contract period, we have been able to obtain very precisestatic and dynamic light scatterino data over the high K2 range (K 2 lxlr I 0

cm 2; 45' :) < 120') for somewhat larger vesicles, and have established that

both Rh and R are consistent (within a few A) with a spherical shell of outer

diameter 580A and thickness 37 A, the latter of which is the generally acteIt-

ed value for lipid bilayers. The point of the latter remark is that we have

now some hard evidence, for at least one preparation, that the actual shape

of the ordinary vesicle species is close to spherical, as is widely believed.

34

Slow secular changes in the properties of all samples (including the

nonirradiated controls) with time, required that the samples and controls be

exposed and analyzed the very same day.

Resul ts

The criterion for assessing damage to the vesicles was whether the

D vs K2 plots for the irradiated samples changed significantly in any respect

when compared with their respective controls, which had not been subjected to

picosecond laser radiation but otherwise had received identical handling, in-

cluding loading into the irradiation cell. Both were observed at the same

time.

Two regions of the D vs K2 curve are of special interest. The high-angle

45 ), or large K 2 , D-values give an indication of the z-average inverse2hydrodynamic radii of the ordinary vesicles. The low-angle, or small K

D-values contain primarily information about scattering centers of larger

dimension, such as clumps of vesicles or large lamellar aggregates (liposomes)

that scatter light predominantly in the forward direction.

Exposure of the vesicle suspensions to entire pulse trains with energy

densities of 600 mJ/cm 2 (peak electric field i,5.8x10 5 V/cm 2) or greater did

produce a distinct change in the D vs K2 curve, as indicated in Figure 2--

.hich shows that the apparent diffusion coefficient decreased by more than 25"

at . = 25 and by a smaller but still significant amouit at higher angles.

Our interpretation of this result is that the number and/or dimensions of the

nonspherical aggregates. presumably multilamellar liposomes, have increased at

the expense of the smaller ordinary vesicles destroyed by the light pulse.

Data for the various samples are summarized in a general way in Table 4.

Values of the apparent diffusion coefficient observed at various angles are

recorded in Table 5.

The massive damage sustained in sample 1, which was exposed to an entire

pulse train, completely overshadowed the more modest changes in sample 2,

which was exposed t" a single pulse. However, after observing the behavior

of the remaining samples, which include some samples of clear-cut negative

results as well as some significant examples of more modest damage, including

sample 3 (shown in Fig. 2), it seems possible that sample 2 actually sustained

35

1 f (Al 0 S-- 0 0 0

0 0LIL S..

0) C O 00 0)0 0i ICO <r - COj r- -1- 00 o c

C-)

U

m- Co 00 C,

C L

Cj C) C') LO m ". r- m o C C:) LO C'CLI- C C) 00 0 . i 1j C C- C 00 LA LA LA CD

In-'

COL)

U<- M 0

LO " 1. 0 N-Nto A ~ A L A L

LD .- 4 C\I C') X)CJ 0 C) ' C D (' V '

10- N- CoUo->- : o o r- t- Lo m L

-O LU ('O C') LA \ C (1

LU c

0..~1 0- 4-) 4" CO C 04AC N- (4 (

cl 4-A LA: 4- 4 ('U 4-) k CM r- ) 00) mV 4-3

LU)

-36

2some significant damage after all. The pulse energy of 9 mJ/cm in that

single pulse corresponded to a peak electric-field strength (assumed uniform

over a lO-psec pulse) of %7.1xIO5 V/cm.

Curiously, sample 5, which was exposed to an entire pulse train at very

high energy density, showed rather slight indication of damage in the sense

that there was a decrease in [) only dt small K2. none at 'large K2 . However,

that difference between the irradiated sample and control persisted into the

second day after exposure, as indicated in Table 6, thereby increasing our

confidence in that result. That sample was irradiated in the longer, narrower

bore cell to achieve the higher energy density. These vesicles are known to

stick to glass, and the large increase in surface area may have "fractionated"

the irradiated sample in an unknown way.

TABLE 6. APPARENT DIFFUSION COEFFICIENTS OF DPPC VESICLESAS FUNCTIONS OF SAMPLE AGE

(D x lO8 cm2/sec)

Sample A e = 250 = 900 = 1200 Damage

3 day 1 4.53 7.29 yesday 2 1.49 6.90

4 day I 5.82 7.55 noday 2 4.97 7.22

5 day 1 1.40 5.07 slightday 2 0.932 4.72

Control day 1 2.09 5.01for 5 day 2 1.21 5.12

These results indicate an approximate damage threshold uf about 600 mJ/cm 2

for DPPC vesicles exposed to trains of picosecond pulses of 1060-nm radiation.

In addition, there is some indication that more modest damage might be inflict-2ed by single picosecond pulses with an energy density of 9 mJ/cm . It is

possible that electrostrictive or other forces present during the pulse actual-

ly rupture the vesicle every time, but that substantial recovery occurs be-

tween pulses in these systems. Certainly, vesicle fragments do not move far

in 10-9 sec, and lipids are insoluble in aqueous solution. Thus, the observed

damage may be minimal unless the vesicle is subjected to repeated battering

by closely spaced pulses, as in a pulse train. In a system where chemical

potential gradients and osmotic pressure differences exist across membrane, as

37

6 - -.. . . . . . . . .. . . . .. . ... -. .. 1

in living systems, a single rupture might lead to complete destruction of the

osmotically strained vesicle, so the damage threshold in terms of peak elec-

tric field might be the same for single pulses and whole pulse trains in that

case. Indeed, our studies of damage thresholds in primate retinae support

this hypothesis. (See Part II of this report.)

It is also evident from Table 6 that the differences between samples 3

(damaged) and 4 (undamaged) persist into the second day after exposure, despite

the long-term changes in both samples. Again, this argues that the differences

between these samples are genuine.

The possibility of a thermal damage mechanism playing a role in the above

experiments can be unequivocally ruled out--the temperature rise in our samples,

as estimated from the difference in energy between the incident and the trans-

mitted pulse trains, was less than 0.02"C.

Egg-Yolk Lecithin (EYL) Vesicles

Our results for DPPC vesicles apply to the ordered low-T phase of themembranes prevailing below Tm 410'C. The question naturally arises whether

similar results would be found for membranes comprised of natural lipids, such

as egg-yolk lecithin (EYL). Although both lipids are phosphatidyl cholines,

the DPPC has two identical saturated 18-carbon fatty acid chains esterified to

the glyceryl moiety, whereas EYL phosphatidyl cholines generally contain a

saturated fatty acid esterified at the 1 position in addition to the unsatu-

rated fatty acid (usually oleic) at the 2 position. As a result of this chem-

ical difference, and also the mixture of phosphatidyl cholines present, the

EYL membranes exhibit a broad thermal transition between -15' and -7'C and are

believed to exist in the fluid phase at room temperature, v20°C.

Preparation of EYL Vesicles

Considerable experimentation with parameters governing the vesicle

preparations was conducted. The effects of mixing-temperature, initial con-

centration, needle gauge, and injection speed were all investigated. The

D vs K2 curves for some of the resulting suspensions are indicated in Figure 3.

The firal procedure is described below.

38

6.0 - I V

V5.0 c

III

n 4.0

C\J I

3.0-

2.0-

1.010 1.0 2.0 3.0 4.0 5.0

K2 x10-10cm-2

Figure 3. D vs. K2 for different preparations of EYL vesicles.

Prep. I. Step 1 (see text for steps) of the procedure was modifiedas follows: Inject 0.94 ml of 10 mg/ml stock into warm(40'C) buffer. Step 3 is omitted.

Prep. II. Step 2 was modified by injecting into warm (400C) buffer.Step 3 was omitted.

Prep. III. Step 2 was modified by injecting into warm (40'C) buffer.

Prep. IV. Step 2 was modified by using 0.94 ml of 10 mg/ml stock.Prep. V. Same as step 4.

39

For EYL commercially supplied in ethanolic solution, the following steps

are carried out:

(1) Rinse out a 25-mil flask and flush with filtered (0.22-um Millipore)buffer, leaving the magnetic stirring rod in the buffer solution tobe rinsed at the same time.

(2) Withdraw 0.5 of 10 mg/ml stock solution (95% ethanol), allow it towarm up for several minutes in syringe, then slowly inject into a25-mil flask containing 10 ml of filtered (0.22-pm Millipore) buffer.Use a Hamilton 1-cc syringe with a 22-gauge needle. The concentra-tion in the flask is now 0.47 mg/ml.

(3) Filter the suspension through 3.0-vm Millipore filters twice.

(4) Clean out scattering vessel with 0.22-1.m filtered buffer. Transfer0.45 ml of solution into cell and dilute with 9 ml of clean buffer.Final concentration is 0.22 mg/ml lipid.

For lyophilized EYL make a 10 mg/ml solution in 95' ethanol first, then

follow steps 1-4 above.

The buffer used for all experiments was 0.1 M NaCl + 0.01 M Tris (Sigma

Trizma-base MW 121.1) titrated with HCI to pH 7.0.2 2

The D vs K2 curves in Figure 3 show no limiting plateau at high K , which

indicates a greater degree of polydispersity than found for DPPC vesicles.

Damage Studies of EYL Vesicles

EYL vesicles were expcsed to mode-locked 1060-nm laser pulse trains with

energy densities between 400 and 660 mJ/cm2 . The results of dynamic light

scattering measurements on the irradiated samples and corresponding identical-

ly treated controls are presented in Table 7. Although samples 6, 7, and 8,

irradiated at the higher energy densities (657, 652, and 614 mJ/cm2 respective-

ly), exhibit D values marginally smaller than those of the controls at every

angle, the data are not sufficiently different to warrant a definitive conclu-

sion that damage occurred.

Unfortunately, all exposures at energy densities exceeding 600 mJ/cm2

were obtained by decreasing the beam diameter by 20%, so not all of the sample

solution was exposed to the full power of the irradiating beam. This practice

was necessitated by the comparatively low energy output of the laser even with

all operating parameters optimized.

Although the D vs K2 curves differed only marginally from the controls

for the most intensely exposed samples, it is still possible that greater

differences were manifested in the time-course of the decay of the correlatinn

a 0

TABLE 7. VALUES OF D AT VARIOUS K2 FOR EYL VESICLES

DxIO8 (cm/sec)

E K2= K2 = K2 = K2 = K2

Sample (mJ/cm 2 0.47x1010 1.03x10 10 1.75x10 10 3.51x1010 5.26x1010

1 400 3.68 4.32 4.89 5.72 6.142 420 3.51 4.03 5.07 5.79 6.18

Control 3.92 4.40 5.11 5.81 6.22

3 420 3.39 4.40 4.97 5.45 5.914 516 3.05 4.17 4.94 5.48 5.88

Control 3.51 4.16 4.73 5.60 6.03

5 434 3.20 3.89 4.49 5.03 5.486 651 3.10 3.82 4.35 4.88 5.29

Control 3.18 3.98 4.39 5.09 5.46

7 652 3.15 4.01 4.52 5.01 5.438 614 3.02 3.80 4.37 4.90 5.27

Control 3.27 4.09 4.63 5.07 5.57

function. That is, the quality of single-exponential fit may still have

varied significantly between irradiated samples and controls. A method for

quantitatively measuring this deviation from single-exponential decay is

that of cumulant analysis.

The normalized intensity autocorrelation function g(2) (T) is a 4th-order

correlation in the scattered electric field. Under the assumption of a

Gaussianly distributed scattered field, g(2) (1) may be related to the 2nd-

order correlation function g(1)(T) by (52)

9(2)T ) = 1 + Bjg(1) ()I2 (32)

where is of order 1. If, furthermore, the sample is polydisperse, g(1)(T)

is weighted by the relative intensity distribution of scatterereG(r) with

different decay constants K 12D,

Ig(lT)j : G(I')e- dr <e- > . (33)

The exponential may be written in the form

e =e e-(v-

41

where

f £ J G(r)i' dr (34)

0

The factor ei - )T may be expanded and inserted in Eq. 33 with the following

result:

S2 3 (5g(1)( ) = e(1+ 2T ~ 3 "'+"* (35)

where

112 - G(r)(r-f) 2dF

u3 { g(r)(p-f) 3 dr

Thus, the intensity autocorrelation function may be expressed in terms of the

moments of G(i'). It is then easily seen that (52)

1 (2 2T-71 n[g)T n f - T,r + !-*2 3 + --- (36)

-] 2 - T+2!L 1 P

where use has been made of the expansion

ln(1+z) z-z2/2 +-

Equation 36 is a polynomial in time whose coefficients are the moments of

the distribution G('). The polydispersity of a sample of scatterers is here

defined as the ratio of the second moment to the square of the average, that

is,-2

polydispersi ty 2/

A program was written to analyze certain sample runs according to the

above polynomial. The algorithm first subtracts a baseline from the raw auto-

correlation data. This baseline is first taken to be either the average of

4?

the last 25 points or the baseline resulting from a single-exponential fit.

Then the algorithm takes the natural logarithm, does a least-squares fit to2the polynomial, and calculates x , the sum of the residuals, which is compared

2 2with the previous X . The baseline is varied until a minimum in , is obtain-

ed. The corresponding values of 2F, u2, (2F)-1 , and polydispersity ( 2/ 2

are printed on the teletype.

The polydispersities obtained for correlation functions at small scatter-

ing angles (e=30" ) always give higher values than at o=120', as shown in

Table 8. The high-angle polydispersities were generally quite similar (i.e.,

within 10') for the irradiated samples and their respective controls, whereas

the polydispersities at 0=30" usually were substantially larger (by a factor

1.3) for the irradiated samples than for controls, as also shown in Table 0.

Thus, the cumulant analysis indicates an appreciable change in the irradiated

sarlples studied (at 435 mJ/cm 2 and 650 mJ/cm 2 ) with respect to their controls.

Unfortunately, this change was not consistently manifested. Moreover, cumu-

lant analysis has some inherent weaknesses. The final results are inordinately2

sensitive to the value of the subtracted baseline, because the minimum in Y

with respect to baseline is not always unique. Moreover, as the value of the

baseline was varied, the number of data points to be fitted also changed to

prevent a zero or negative argument of the logarithm. This latter feature had

the consequence that the fitting algorithm in some cases simply slid progres-

sively to higher baselines with ever-fewer data points without converging to2

any local minimum in Y2

TABLE 8. POLYDISPERSITY / at o = 300 AND e 120'FOR EYL2VESICLES

Sample 5 Sample 6

435 mJ/cm2 650 mJ/cm 2 Control

o = 30''

K2 - 0.47x10 10 0.233 0.200 0.127

= 1200K2 = 531010 0.125 0.117 0.130

Ratio

(30/120) 1.86 1.71 1.02

43

A.

Summuary of EYL Results

The LYL lecithin vesicles proved considerably more intractable than DPPC

vesicles in every regard from preparation to analysis of their dynamic light

scattering autocorrelation functions. Although some evidence of optical

stress-induced changes, especially in the polydispersity parameter, was ob-

served in some preparations, we feel that the evidence overall is too slim

to warrant at this time an unequivocal statement concerning damage of these

vesicles.

Future Vesicle Work

The next logical step is to examine the susceptibility to picosecond

laser pulses of osmotically strained vesicles; for example, containing concen-

trations of salt inside differing from those prevailing outside. In that

case, which corresponds more closely to living organisms, one has the possi-

bility that a single rupture event will be amplified to complete vesicle

destruction by the osmotic gradient, with less likelihood of recovery. In

such a case the damage threshold for single pulses could well approximate

1/N of that for entire pulse trains, where N is the total number of pulses

in the train. In other words, the first pulse to reach the threshold field

would destroy the vesicles, and all subsequent pulses in the train would

simply add 'insult to injury."

44

PART II: PICOSECOND OCULAR DAMAGE STUDIES ON PRIMATES

INTRODUCTION

This phase of the program was designed to determine ultrashort pulse-

induced ocular damage thresholds in the primate Macaca fascicularis* at two

different wavelengths, 530 and 265 rm, obtained by frequency upconversion of

the Nd:Glass laser output, and to compare the effects of single pulses and

entire mode-locked pulse trains at the 530-nm wavelength. In the visible

spectrum the cornea, lens, and vitreous are highly transparent and thus

damage is sustained in the chorioretinal region, whereas at the UV wavelength

the incident radiation is absorbed mostly within the corneal epithelium.

The irradiation apparatus used in both the retinal and corneal damage

studies was similar to that used in the macromolecular exposure experiments.

(See Appendix A.) The output of the Nd:Glass laser was frequency-doubled to

generate 530-nm light for the retinal damage studies and frequency-quadrupled

to generate 265-nm radiation for the corneul threshold experiments. Addition-

al details are given in the followinq sect-un.

The experimental protocol, parameters. and criteria we used followed as

closely as possible those used in the bulk of previously published work, in

order to permit direct comparisons with data acquired in prior research by

other investigators.

We describe first the retinal damage threshold studies, which comprise

two separate bodies of experimental data: (1) damage thresholds for entire

mode-locked pulse trains, and (2) damage thresholds for single ultrashort

pulses. The corneal ultraviolet irradiation experiments are discussed in

the final section of this report.

*Orirjinally we planned to use rhesus monkeys (Macaca mulatta); however,the continuing embargo on the exportation of this species by India and there',ulting severe shortage dictated the use of M. fascicularis. The structureand pigmentation of the retinae of the two species are very similar, allowingdirect comparison of the present work with past results.

45

RETINAL DAMAGE THRESHOLDS INDUCED BY PICOSECOND 530-nm LIGHT PULSES

Irradiation Apparatus

The irradiation apparatus used in these studies had two somewhat different

configurations: the first designed for experiments using the entire mode-

locked pulse train of %100 pulses, and the second for work using a single

pulse selected from near the beginning of the pulse train.

Configuration Used for Pulse Train Studies--A schematic of the apparatus

is shown in Figure 4. The infrared pulse train was frequency-doubled to 530 nm

in a Type I phase-matched KD*P crystal angle-tuned for maximum conversion ef-

ficiency (n,12%). The 530-nm component was separated from the 1060-nm light by

a dichroic beamsplitter (DBS). A Schott KG-3 filter blocked residual IR from

the green beam. A half-wave retardation plate (X/2) rotated the polarization

of 2nd harmonic light from the vertical to the horizontal plane. An uncoated

pellicle beamsplitter (PL ) sampled a small fraction (%1.7') of the green beam

for pulse chronometry. The green pulse train was attenuated to approximately

the desired energy level by a neutral density filter stack (NDFI). An Iris

diaphragm (ID) selected the central 3 mm from the green beam, which had di-

verged to a diameter of -i0 mm at the location of this aperture.

The 2nd harmonic pulse energy was monitored by a Laser Precision RkP-331

pyroelectric energy probe calibrated by the manufacturer using standards trace-

able to the National Bureau of Standards. The pulse energy was sampled by an

uncoated fused-silica beamsplitter (BS, total reflection coefficient = 7%,

both faces), and focussed by a 15-cm-f.l. fused-silica lens into the aperture

of the probe. The energy in the pulse train was displayed on a Laser Precision

RkP-3230 digital readout unit.

Additional beam attenuation was provided, when desired, by a set of cali-

brated fused-silica neutral density filters (NDF2).

Configuration Used for Single-Pulse Studies--The setup used for single-pulse

irradiaticn was somewhat different from that dcscribed above. (See Fig. 5.)

First, the Pockels cell pulse-selecting system, installed ahead of the 2nd

harmonic generator, was activated as described in Appendix A. Because of the

low energy per pulse (-,1% of the total energy in the pulse train), the half-

wave retardation plate (X/2) was placed downbeam from pellicle PLI. This

increased the reflection coefficient of the pellicle by about an order of

46

ID KG-3 OBS KD*P SHG 16-i

M + 9lfl PULSE TRAINI FROM NId:GLASS

1060 flfl 530 nm LASER

=3BG-38

q L 2 7,2

CCTV 4 L

STREAK L 4 L 3 NDF 3SHUTTER ND IF

VDC t~lBS

OSCILLOSCOPE NDF 2

PL2

EYE CAMERA

Figure 4. Schematic of apparatus for irradiation of primate eyeswith entire trains of ultrashort 2nd harmonic (530 nm)pulses derived from a mode-locked Nd:Glass laser.

47

ID KG-3 DBS KD*P SHGM - flu [=I SINGLE ULTRASKOR~T

LJ # 1060-nm PULSE1.060 nm 530 nm

X/ 2

L BG-38

L2

r C C T V _ _ _

t STREAK L4 L 3 NDF1 j X/SHUTTER L6

VDC ID

BS

CCD PRB DIPA

OSCILLOSCOPE ND O 2

EYE FUNDUSCAMERA

Figure 5. Schematic of apparatus for irradiation of primate eyeswith single ultrashort 2nd harmonic (530 nm) pulsesderived from a mode-locked Nd:Glass laser.

.1)dy9ni ri~de ,n,_. the oci,:t o)u'se was r ow verticadly polarized. Without thi:

modification not enough light would have entered the pulse chronometer system

to yield a detectable signal.

The most significant change made for the single-pulse studies was the

acdition of a beam-reducing Galilean telescope, comprised of lenses L6 and L7.

The diameter of the beam was reduced by a factor of 3.7 so that all of the

green pulse energy passed through the 3-mm-aperture iris diaphragm. This beam

reduction was necessary so that the single pulses could reliably trigger the

pyroelectric energy probe. Without the beam reduction, only about 10C of the

pulse energy was transmitted by the aperture, resulting in too low a pulse

einergy at the energy probe for proper triggering. Unfortunately, the use of

the beam-reducing telescope increased the beam divergence by a factor equal to

the beam reduction ratio. Thus, the beam divergence, and hence the retinal

irradiation spot size, for the single-pulse experiments was 3.7 times greater

than in the experiments utilizing the entire pulse train. (The method used to

determine the beam divergence is discussed in the section below.) The prob-

lem of poor energy probe triggering on single pulses was not discovered until

after all of the pulse train data had been obtained, so the two sets of experi-

ments could not be carried out using the same beam divergence.

Control of pulse energy incident on the eye for the single-pulse case

was effected by placing calibrated neutral density filters (NDF2) between the

energy-sampling beamsplitter and the eye to be irradiated. The pulse energy

was monitored as described earlier.

Apparatus Common to Both Configurations--The pulse chronometer system,

based on a transverse-gated optical Kerr effect shutter, is identical to that

used in our macromolecular irradiation studies (Appendix A). It provides an

on-line measure of the pulse duration of the IR laser output. The frequency-

doubled pulse duration can be deduced from the fact that its intensity scales

as the square of IR intensity. Thus, assuming that the laser pulses have a

Gaussian temporal shape, the duration of the 530-nm pulse is I1/V that of the

IR pulse. In this work, the IR pulse duration was in the range of 4-13 psec,

with an average of 9 psec. It follows that the 530-nm-pulse durations fell in

approximately the 3-9-psec range, with an average of %6 psec.

The beam spatial profiles were measured by directing the 530-nm beam onto

i lensless Fairchild CCD camera having 1024 channels/inch. The linear CCD

49

diode array was placed at the same distance from the iris-diaphragm aperture

stop as was the pupil of the test primate. The beam profile was approximately

Gaussian in both the single-pulse and entire pulse-train configurations with

a I/e 2 diameter of <4 mm at the pupil in each case. Beam divergence was mea-sured in a similar manner, with the CCD array placed at the focal point of a

50-cm-f.l. lens. The divergence was -6 mrad for the single pulse and "1.6 mrad

for the pulse trains, indicating irradiated spot diameters on the retina of

,78 pin and %21 pm, respectively, based on a typical value of 1.3 cm for the

focal length of the primate eye (53).

The experimental primates were mounted in a prone position on an adjust-

able platform having 3 degrees of translational freedom (x,y,z) and 2 degrees

of angular freedom (azimuth and elevation). The pupil of the subject eye could

be positioned with respect to the laser beam axis by one or more of these

mechanical adjustments. Visual examination of each primate retina was per-

"ormed for t)roper placeinert of the test exoosures :nd 'or pre- and postexposure

examination to determine incurred damage.

A dielectric-coated pellicle beawsplitter (PL2), attached to a two-posi-

tion swingaway mount affixed to the fundus camera objective barrel, directed

the laser pulse into the eye. In the "down" position the beamsplitter reflect-

ed 210.54* of the incident laser energy into the eye, collinearly with the

optic axis of the fundus camera, thus permitting simultaneous observation of

the irradiation event. This on-line observation capability permitted precise

positioning of the irradiated spot, allowing corrections for eye motion right

up to the time the laser was fired. With the beamsplitter in the "up" posi-

tion, the eye could be examined and photographed with no loss of illumination,

or vignetting.

Experimental Protocol

Ten Macaca fascicularis primates were used in the experiments. Prior

to the irradiation studies the primates were refracted in each eye to

-*The 45' reflection coefficient of the dielectric-coated beamsplitterfor horizontally polarized light at 530 nm was determined experimentally,using a collimated CW light source of the correct wavelength and polarization,chopped at a frequency of 330 Hz, and a photomultiplier detector. The re-flectivity was deduced from measurements of the transmission factor, usingthe fact that absorption losses in the coating are loss than 0.2%.

50

) ) the iorrct iv o s ~ up -)I t! i I i o.-d (_Yc 1 plJe. j ( wereobtained by instillation of twu drops of Kupfer's solution, a one-to-one mix-

ture of I cyclopentolate and 10 phenylephrine.) Any eye with a refractive

error greater than 1.50 diopters in any meridian was not used in the experi-

nents.

One day prior to the retinal irradiations of a given primate, pupillary

dilation was initiated by instillation of two drops of atropine sulfate (4"

solution). This was followed, immediately prior to irradiation, by two drops

of Kupfer's solution. Approximately 30 minutes prior to irradiation, each

animal was deeply sedated with an intramuscular injection of 0.6 cc ketamine

HCl. Booster doses of 0.2 cc ketamine were administered during the course

of the experiment as needed. No additional general anesthetics or tranquiliz-

ers were required. Eye movements were eliminated by retrobulbar injections

of xylocaine (0.35 cc each side of the orbit). A subcutaneous injection of

0.15 cc atropine sulfate served to suppress drooling.

During the exposure procedures the eyelids were kept retracted by a pedi-

atric stainless-steel speculum. Corneal dessication was prevented by frequent

irrigation with normal saline, using a modified hypodermic syringe equipped

with a 13-G short cannula aimed at the eyc. The syringe was connected to the

saline solution container through a T-type double ball valve. The syringe

acted as a positive displacement manual pump in this configuration and was

adjusted to deliver 2 cc of saline per stroke of the spring-loaded plunger.

To guide the placement of the irradiation sites, eight marker lesions,

four arranged vertically and four horizontally, were placed adjacent to the

macula, as shown in Figure 6. These markers defined a cartesian coordinate

system for the 16 test exposures within the macula. (In some cases a fifth

row or column was also irradiated for a total of 20 exposure sites within the

macula.) The horizontal marker row was always inferior to the macula, whereas

the vertical row was temporal in the right eye and nasal in the left eye. The

marker lesions were produced with the same laser system used to produce the

experimental lesions, but at considerably higher energy. The entire pulse

train of 530-nm light was used at an incident energy of 30-50 vPJ. The result-

inq lesions appeared immediately in most cases and had a diameter "3 times the

spot. size of the laser beam at the retina. Most of the marker lesions appeared

d , whitish discolorations,, although a few exhibited some subretinal hemorrhag-

iu, probably a result of damaging hidden capillaries.

51

LD '-I >'-II

C)

V)-

C)I

(0

C)C

C14 r- CD LA >

00) LA >

Cross-hairs in the eyepiece of the fundus camera were used to align the

exposure site with the marker lesions. The animal was moved relative to the

laser beam to change the exposure site. The approximate energy range of a

given series of exposures was adjusted by the first neutral density filter

stack (NDFI). (See Fig. 4.) Final adjustments were made with the calibrated

neutral density filters (NDF 2). The laser itself provided a degree of random-

ness in the incident energy as a result ot its fluctuations in output from

shot to shot. The IR output of the laser typically fluctuates over a range

of 25'J of its average output. Since the frequency-doubled energy scales as

the square of the IR output, the 530-nm incident energy fluctuated over a

range of '50', from shot to shot. In cases where a single pulse was selected

from the pulse train, the fluctuation was even greater because of the varying

efficiency of pulse selection by the Pockels cell from shot to shot. (See

Appendix A.) This randomness in the energy delivered to successive exposure

sites was useful in preventing bias when evaluating the postexposure sites.

Results

The maculae of each primate were examined with the fundus camera by two

observers for the presence or absence of visible lesions at I hour and 24 hours

post exposure. P.lesion was defined as the smallest observable circular dis-

coloration (usually whitish or light gray) differing from the retinal back-

ground. Using this criterion, 148 exposures were made in the range of 0.089-

19.5 J with the whole pulse train, and 158 exposures in the range of 0.01-

7.1 pJ with single pulses, to establish the ED50 point (the energy required to

produce a lesion in 50,' of the events) in each case.

Before the data were analyzed, the pulse-train and pulse-width data were

carefully examined to eliminate invalid exposure sites. The criteria for this

test were as follows:

(1) For pulse trains, the train had to be clean; i.e., consist of a

series of '100 equilly spaced pulses. Interlaced multiple-pulse trains or

the presence of spurious pulses between the main pulses were cause for data

rejection.

(2) For sinqle pulses, more than 90 of the energy had to be in one

pulse. On a number of occasions two successive pulses in the train were switched

out by the Pockels cell. Such events were riot used ir the data reduction.

53

Occasionally, spurious pulses spaced a few tens of picoseconds from the select-

ed pulse would appear. Those data shots were also rejected.

In addition, exposure data were rejected if during the actual laser shot

the laser spot geometry deviated appreciably from a circular geometry or if it

appeared fuzzy and significantly larger than normal. (Such effects could be

caused by gross refractive errors resulting from distortion of the cornea by

the speculum or by side effects of the retrobulbar injections.) Using this

criterion alone, all data from three eyes had to be rejected.

In the case of the whole pulse train experiments, 77 exposure events passed

the above criteria. With the single-pulse experiments, 108 events passed.

These numbers were sufficient to establish reliable ED50 thresholds for each

category.

The data were submitted to standard statistical probit analyses (54) to

determine the ED50 points (55). The analyses were carried out on all exposure

sites in all eyes exposed to a given set of conditions. Implicit in this com-

bined-probit approach is the assumption that the variability from eye to eye is

no greater than the variability among sites within a given macula. This assump-

tion has been demonstrated by others to be valid (23). The calculations, how-

ever, were carried out separately for the lesion/no-lesion data of each of the

two observers. Table 9 presents the I-hour and 24-hour ED50 points for each

observer, with the associated 95% coifidence ranges noted in parentheses.

TABLE 9. RETINAL DAMAGE THRESHOLDS AT 530 nm

Energy density at retina*Pulse Single Evaluation (Average of both observers)

train pulse time Obs. A Obs. B ( 2/cm2)X 1 hr 4.22 oJ 4.46 uJ 1.1

(3.64-4.91) (3.76-5.29)

X 24 hr 1.89 0J 2.31 Ji 0.54(1.49-2.38) (1.87-2.84)

X I hr 0.61 ;!J 1.03 iJ 1.5x1O -

(0.38-0.98) (0.65-1.64)

X 24 hr 0.24 jIJ 0.24 J 4.4x10 -"

(0.17-0.35) (0.15-0.39)

These data corrected for transmission factor of 0.88 of clear ocularvedia at 530 nm (29).

51

The ED50 thresholds for single-pulse le,ioris are considerably lower thanthe thresholds obtained with the entire pulse train. This is especially evi-

dent in the 24-hour results in which there is apparently a one-order-of-magni-

tude difference between the single-pulse and whole pulse-train thresholds. The

difference is actually much greater when we consider the fact that the retinal

spot size for the single-pulse case (,,78 0m) is *3.7 times greater than for the

pulse-train case (%21 om). Thus the energy density at the retina is an addi-

tional factor of 13.7 lower for the single-pulse case. In other words, the

actual threshold for damage for the single-pulse case is about two orders of

magnitude below that for an entire pulse train, in terms of the energy density

deposited at the retina, as is evident in the last column of Table 9.

An interesting deduction that can be made based on our results is that,

for the case of the pulse train, the lesion event may actually be caused by

the first pulse that reaches the threshold energy for a single pulse. General-

ly this occurs near the beginning of the train. The reasoning is simple:

There are I00 pulses in the train and thus the energy per pulse is I/100 of

the total, which is approximately the same factor by which the single-pulse

threshold is smaller than the pulse-train threshold in terms of energy density

at the retina. The remaining pulses in the train thus would seem to only add

"insult to injury." This also suggests that, if no pulse in the train meets

the threshold energy value for single pulses, there may be no damage regardless

of the number of pulses in the train, as long as the total pulse-train duration

is less than a few microseconds. For longer trains, thermal effects may begin

to come into play and this hypothesis would be invalid.

It is instructive to compare the retinal damage thresholds at 530 nm to

the res ults of our earlier vesicle damage studies at 1060 nim, in terms of inci-

(lent energy density, power density, and corresponding electric-field strength.

Table 10 ,hows thdt the 24-hr retinal thresholds are remarkably close to the

vesicle thresholds for both the pulse train and single pulses, also that the

da,3,age threshold oi , sin(Ile )ulse is about two orders of magnitude below

fhoi fr i, eritir, 1,ul ,,, dt in )f 1OU pulses for both the vesicles and the

r: ' K , ' I' n rldi( ation that the damage mechanism in both cases1 ,',f, ,' h ,. ,k i-r)wer density, and hence the peak electric field in

r in Ir ~ ,1j~' , ,, ther than on the total deposited energy density. The

fid, ,u1* .',',ce to the hypothesis forwarded earlier that the

55

first pulse in the train to reach or exceed the single-pulse threshold wiil

cause the damage. Subsequent pulses in the train would increase the number

of damaged vesicles or cells but would not be necessary for a detectable chani

or lesion.

TABLE 10. COMPARISON OF THRESHOLD DAMAGE DATA FOR VESICLES AND RETINAS

Vesicle Damage Thresholds Retinal Damage Thresholds*1 1060 nrn At - 10 psec) = 530 nin, At - 6 psec)

Peak elec- Peak elec-Energy Power tric field Energy Power tric fieldJ/cm 2 GW/cm 2 V/cm x 105 J/cm2 GW/cm 2 V/cm x 105

Pulse 0.6 0.6 5.8 1 hr 1.1 1.8 10train 24 hr 0.54 0.90 7.1

Single 9x1O-3 0.9 7.1 1 hr 1.5x1O-2 2.5 13pulse 24 hr 4.4x10-3 0.73 6.4

*Retinal damage thresholds are corrected for the transmission factor ofthe clear ocular media at 530 nm. T5 30 = 0.88 (29).

Since the electric fields associated with the threshold values are nearly

an order of magnitude greater than the membrane potentials of both the vesicle

and cell lipid bilayer membranes, it appears reasonable at this time to ascribe

the ocular damage to membrane disruption by electrostrictive forces.

This tentative identification of the damage mechanism must be tempered

somewhat, at least at present, by a number of factors. First, the wavelengths

used in the vesicle and retinal work are not the same. Thus, photobiological

processes may be present in the retinal case and the damage may not be a result

of membrane disruption by electric fields. Unfortunately, it was not possible

to generate sufficient power in the 2nd harmonic of the Nd:Glass laser t,.

attain the 1060-nm threshold levels for the vesicles at 530 nm. Such an exrc -

ment would require a laser amplifier of gain 3, which was not available to us

at the time of the research. Thus, the wavelength dependence of vesicle Jii&a'

could not be determined.

Second, the vesicle data are based on relatively few individual expevi-

ments, so their damage threshold determination is statistically less reliable

than for the retinae. The reason for the relatively few data points is that

the preparation of the vesicle samples, their irradiation, and their s,, ,

5

examination by dynamic 'ight scatterinq are very time consuming. X,

more than one or two samples a week could be studied in this manner, whereas

all of the retinal data (320 data points) were gathered in a matter of a few

weeks.

Finally, the vesicles we studied were not subject to osmotic pressure

imbalances or chemical potential gradients across the membrane, whereas such

effects are present in living cells. In the living cell the disruption of the

membrane by the electric field in a single threshold laser pulse would be

facilitated by the osmotic pressure imbalance across the membrane, resultingin irreversible damage.

Comparison with Other Work

The 24-hr postexposure results reported here are compared in Table 11

with the results of similar experiments carried out by Ham et al. at 1064 nm

(28), Goldman et al. at both 532 nm and 1064 nm (29), and Taboada et al. at

1060 nm (30,31). A considerable discrepancy appears not only in the ED50values but also in the threshold power density and electric field at the

retina. In addition to differences in wavelength and pulse duration, a number

of possible factors related to experimental protocol and the definition of a

threshold lesion can significantly affect the experimental results.

One factor may be that in none of the work is the diameter of the irradi-

ated spot size at the retina known with a great degree of certainty. This is

a difficult parameter to measure in vivo (the diameter of the lesion is not

necessarily the same as that of the irradiated spot); consequently, the spot

size is usually estimated from a knowledge of the laser beam divergence and

an estimate of the effective focal length of the emmetropic primate eye.

Although all researchers, including ourselves, have used corrective

lenses if the refractive error of the primate eye exceeded ±0.5 diopter in any

plane, nonintrinsic refractive errors can arise during the irradiation experi-

ments. On a number of occasions, for example, we have observed otherwise

normal eyes become highly astiqmatic during an experiment, possibly as a re-

sult of delayed adverse reaction to the retrobulbar injections or, to a lesser

extent, in response to the forces exerted by the eyelid retracting speculum.

On other orcasinns wp also noted that the visibility of the retina

through the fundus camera wa , greatly reduced due to opacification of the

57

a) .4 - i C. CD -I .

4--) C- .) C ) ~ M

Q) LU

4-) C D10 M

V) -0 C

ZD C)L/) 00

I-J LO Zd D 0I0 1_D 0: IT~ C) CQ I') C

CM >l- 4-- a) -_j >< Ej 4-) 4)'

u-iL E C) r- c) C COj L.) CM

C))

(AC) C)4-' C) ~

'-4 Cj rn Li' 00' C) C)(d-4-)

tC) L t

LI

Cl)C a)i k1 CM .0 C) ( CO) C)

C)

<LW F(Cfl C") a) ) C") 3 a

(DL->t0o C) c uC)C- L. - - .- S

4- L) V) C) (n 4-- (A

rCC)D CD C -: D C) C

IC) L Lo' C) CD LO' C) CD

a))

4-) 4-) 4m)

a) C) )Ln

CCL C (

cornea. Frequent irrigation with normal saline did not always improve

visibility.

The above factors can lead to considerable error in the estimation of

the irradiated spot size and to significant attenuation of the incident beam

due to corneal scattering. In either case, the radiant exposure at the retina

could be significantly lower than expected, resulting in high values of ED50.

The experiments reported here are the only ones in the picosecond regime

in which the fundus was observed during actual firing of the laser. The

laser was fired only under conditions of maximum clarity of the ocular media.

If any gross astigmatism was evident at any time during an experiment, the

data for that eye were rejected. Not all previous work has followed such pre-

cautions.

Other variances in exposure protocol and, particularly, variances in

threshold determination can also significantly affect the experimental results.

In two cases (28,29) the definition of a threshold lesion was one which became

just visible fundoscopically within 24 hr of exposure. No statistical analyses

of the data were undertaken by those workers, so the determination of what con-

stituted a threshold lesion was very sensitive to the observer's subjective

interpretation. On the other hand, the work of Taboada et al. (30,31) and the

results presented here made use of probit analysis to determine the damage

thresholds and are thus less prone to subjective errors.

CORNEAL DAMAGE THRESHOLDS INDUCED BY PICOSECOND 265-nm LIGHT PULSES

Irradiation Apparatus

The facility used in the UV corneal irradiation experiments was in most

respects similar to that used in the 530-nm retinal exposure experiments. A

schematic of the apparatus is shown in Figure 7. When compared with the facil-

ity shown in Figure 4, the UV facility differed in two respects: (1) a temper-

ature phase-matched KD*P crystal was added to frequency-double the 530-nm light

to 265 nm; (2) the primate was placed so that the UV beam impinged on the

cornea directly, without deflection by a beamsplitter or mirror.

In addition, a pair of Schott UG-5 UV bandpass filters, having a combined

transmission of 50 at 265 nm and 10-6 at 530 nm, was placed at the output of

the 4th harmonic crystal to eliminate the 530-nm light. A 20-cm-f.l. fused

59

_0!

DBS KD*P SHG1060 nm 1060-nm PULSE

TRAIN FROMNd:GLASS LASER

TOSTREAK BG-38

SHUTTER X/2

PL1

Li

rq L2

KD*P FHG

265 nm

-- -ID

BSL3 ENERGY DITA

PROBE DIGITALPROBE DISPLAYNOF

PRIMATE EYE

Figure 7. Schematic of apparatus for irradiation of primate eyeswith ultrashort 4th harmonic (265 nm) pulse trainsderived from a mode-locked Nd:Glass laser.

60

quartz (Suprasil) lens (L4) was used to vary the beav, diameter impinginq on

the cornea to vary the incident intensity over the desired ranqe. A set of

calibrated UV neutral density filters (NDF) was used for additional control

of intensity. All neutral filters ahead of the UV generator were reilloved to

assure maximum conversion efficiency. In this same regard, the beam diameter

of the 530-nm light was reduced by a factor of 3.7 by lens pair L - L2 to in-

crease the intensity and thus the UV conversion efficiency in the FHG crystal.

The aperture stop was set at 3 mm as in the case of the 530-nm experiments.

The beam diameter at the cornea was set by varying the distance between

the cornea and the focusing lens (L4 ). Maximum diameter was obtained by re-

moving the lens. The cornea-to-lens separations used in the experiments were

10, 15, and 20 cm. Other values of incident energy density were obtained by

means of the calibrated UV neutral density filters. The actual beam spot size

at the cornea was determined by a fluorescence technique. At the various loca-

tions of the cornea relative to the focusing lens, a strip of Scotch "Magic"

transparent tape coated with an ethanolic solution of coumarin dye (Eastman

X5419) was placed perpendicular to the UV beam. The very thin coating of dye

left on the tape where the alcohol evaporated, fluoresced strongly without any

blooming. A closed-circuit TV camera was focused on the rear surface of the

tape to record the fluorescent spot. The intensity distribution in this

spot was determined by the same video analyzing system used to measure pulse

widths with the optical Kerr shutter. Care was taken to eliminate any stray

visible or IR light and to assure that the fluorescence response was linear,

the latter verified by means of UV neutral density filters. The spot inten-

sity profiles determined in this manner closely followed the Gaussian profile

with ie 2 diameters of 2.9, 1.6, 1.1, and 0.76 mm, corresponding to no lens,

and distances of 10, 15, and 20 cm from the focusing lens. These beam sizes

were used to calculate the incident enerqy densities at the irradiated sites.

Pulse duration measurements were carried out as described in Appendix A.

Since the 4th harmonic generation scales closely as the fourth power of the

laser output, the UV pulse durations are half the pulse durations of the 1060-nm

fundamental. Thus, in this set of experiments, the average duration of each

pulse in the UV pulse train was ,4.5 psec.

61

Experimental Protocol

The same ten primates (Macaca fascicularis) used in the 530-nm retinal

irradiation experiments were used in the UV corneal irradiation experiments.

Prior to exposure each primate was rrepared as described previously for te

530-nm exposures, with the exception that retrobulbar injections were not

necessary here. The corneas of each animal were examined with a hand-held

ophthalmoscope prior to the exposure experiment to locate any preexisting

corneal lesions. Only one eye had a preexisting lesion and its character

was recognizably different from laser-induced lesions. Four sites were exposcd

on each cornea, as shown in Figure 8. Only entire pulse trains were used in

this study, due to the relatively low UV energy available; the single-pulse

energy was too low to trigger the energy probe reliably. A total of 80 sites2.were exposed over an incident energy density range of 1.3 to 70 mJ/cm , using

the method described in the Irradiation Apparatus section to vary the energy

density.

Results and Discussion

The corneas were carefully examined with an ophthalmoscope immediately

after exposure, 1 hr post exposure, and 24 hr post exposure. In addition, at

24 hr post exposure, the corneas were examined with a Nikon slit-lamp bio-

microscope. This examination was done both in white light and with cobalt

blue light. In the latter case, sodium fluorescein dye was introduced into

the eye by the standard technique of inserting a dye-impregnated paper strip

under the lower eyelid and then spreading the released dye over the cornea by

blinking the eyelids manually. The sodium fluorescein adheres preferentially

to the corneal lesions and glows greenish yellow under cobalt blue illumina-

tion. This was the most sensitive test for the existence of threshold lesior,4.

The results were recorded photographically as well as visually. No lesions

were detected either immediately or 1 hr post exposure, even at the highest

energy densities. However, epithelial lesions were observed in about Idif

the cases at 24 hr post exposure. Under white-light illumination, the lesi(,t.

appeared as small circular grainy opacifications, having a diameter somewhat

smaller than the UV beam diameter. None of the lesions appeared to have ary

significant depth. Stained with sodium fluorescein and observed under cobal;

62

CD)

CL 4A

-)

4A

4--

4

C114

V 0)cmA

C0)

4-)

63)

blue illumination, the lesions stood out more clearly. In all cases the

lesion/no-lesion events were recorded by two observers.

The lenses of eyes that exhibited corneal lesions were also examined with

the slit-lamp for the possible presence of UV-induced cataracts. However,

unlike work reported at longer wavelengths, >325 nm (26,27), no signs of any

lenticular opacification were seen in any of the eyes. Retinal examinations

were not made since the subject animals had been previously used in the 530-nm

retinal damage studies.

The data were statistically analyzed using the method of probits, as for

the 530-nm retinal exposures. An ED50 threshold was definable only 24 hr post

exposure, as no lesions were observed 1 hr post exposure. Aqain, only data

that represented well-defined operation of the laser were used in the analysis.

(A total of 61 events were usable.) The mean of observer values averaged over

the eyes gave the following estimate with the associated 95% confidence limits:

ED5 0 = 8.2 mJ/cm 2 (6.3 - 10.7) at 24 hr post exposure. The corresponding peak

Power per pulse in the train at the cornea was 18 MW/cm , and the electric

field per pulse was -.105 V/cm.

There are no data on UV laser effects to which these results can be com-

pared directly. The shortest UV laser wavelength previously studied for

corneal damage effects was 337 nm from a pulsed nitrogen laser producing 10-nsec2

pulse durations (27). In that work the damage threshold was 8.7 J/cm , or

three orders of magnitude higher than the present value. No correlations can

be made with that result because of the vastly different pulse durations and

because corneal tissue, as all biological tissue, exhibits a considerable

variation in its adsorption spectrum in the region between 200 nm and 400 nm

(2,3).

The only data available at a wavelength close to the 265 nm studied here

are those of Pitts et al. (25) at 270 nm for a noncoherent UV source (an arc

monochromator) and essentially CW exposure. Pitts and his coworkers found

that the peak sensitivity of the corner to UV photokeratitis was at 270 nm for

both the primate and human eye, and the: determined the threshold to he

4 m]/cm for both subjects. This value is comparable in magnitude to our

results at 265 nn,; however, except for some runrltate Itaininq, the experi-

ments of Pitts did not produce well-defined corneal opacities similar to ou!-.

641

The primary damage mechanism for picosecond pulses at 26, nn! is som-what

speculative at this time, but it appears to be photochornical in hatire.

Neither the power density nor electric field per pulse in the mode-locked

pulse train appears to be sufficient for nonlinear or direct electric-field

effects. For example, the peak electric field of .10 V/cm carried by the

individual pulses is too close in magnitude to the electric field within the

cell membranes (O.6x1O5 V/cm) to have a significant effect.

Although the mechanism of the effect of UV radiation on biological sys-

tems is only generally understood, we believe that the most likely damage

mechanisms at 265 nm are the photochemical alteration, such as denaturation

and coagulation, of proteins and nucleic acids. Especially vulnerable are

unconjugated nucleoproteins of the cell and the DNA in the chromosomes. Both

are particularly susceptible, with maximum sensitivity at 265 nm, since this

wavelength lies at or near the center of their action spectra (2,3). Further

work needs to be carried out at 265 nm in a manner similar to our experiments

in the visible and near IR to confirm this damage mechanism.

65

REFERENCES

1. Sliney, D.H. The development of laser safety criteria--biological con-siderations. In M.R. Wolbarsht (ed.). Laser applications in medicineand biology, Vol. I. New York: Plenum Press, 1971.

2. Michaelson, S.M. Human exposure to nonionizing radiant energy--potentialhazards and safety standards. Proc IEEE 60:389 (1972).

3. Sliney, D.H., and B.C. Freasier. Evaluation of optical radiationhazards. Appl Opt 12:1 (1973).

4. Goldman, L., D.W. Fradin, N. Bloembergen, and D.F. Richfield. Studies inlaser safety of new high-output systems. 1. Picosecond impacts. OptLaser Tech 5:11 (1973).

5. Goldman, L., E. Yablonovitch, N. Bloembergen, and D. Richfield. Studiesin laser safety of new high-output systems. 2. TEA CO2 laser impacts.Opt Laser Tech 5:58 (1973).

6. Mainster, M.A., T.J. White, and R.G. Allen. Spectral dependence of retinaldamage produced by intense light sources. J Opt Soc Am 60:848 (1970).

7. Sliney, D.H., et al. Laser hazards bibliography. U.S. Army EnvironmentalHygiene Agency, Aberdeen Proving Ground, Md., May 1975.

8. Dunsky, I.L., and P.W. Lappin. Evaluation of retinal thresholds for CWlaser radiation. Vision Res 11:733 (1971).

9. Bresnick, G.H., et al. Ocular effects of argon laser radiation. InvestOphthalmol 9:901 (1970).

10. Ham, W.T., et al. Helium-neon laser in the rhesus monkey. Arch Opthalmol84:798 (1970).

11. Frisch, G.D., E.S. Beatrice, and R.C. Holsen. Comparative study of theargon and ruby retinal damage thresholds. Invest Ophthalmol 10:9i1(1971).

12. Vassiliadis, A., H.C. Zweng, N.A. Peppers, anu R.R. Peabody. Thresholdsof laser eye hazards. Arch Env Health 20:161 (1970).

3. Ebbers, R.W. Retinal effects from multiple-pulse galliwi, arsenide rSAM-TR-72-25, Nov 1972.

14. Hayes, J.R.. and M.i.. Wolbarsht. Thermal model for retinal damage inducedby pulsed lasers. Aerosp Med 39:474 (1968).

56

15. Lappin, P.W., and P.S. Coogan. Relative sensitivity of various areasof the retina to laser radiation. Arch Ophthalmol 84:350 (1970).

16. Gibbons, W.D., and D.E. Egbert. Ocular damage thresholds for repetitive-pulse laser exposures. SAM-TR-74-1, Feb 1974.

17. King, R.G., and W.J. Geeraets. The effect of Q-switched ruby laser onretinal pigment epithelium in vitro. Acta Ophthalmnl 46:617 (1968).

18. Ebbers, R.W., and I.L. Dunsky. Retinal damage thresholds for multiple-pulse lasers. Aerosp Med 44:317 (1973).

19. Adams, D.O., D.J. Lund, and P.O. Shawaluk. The nature of chorioretinallesions produced by the gallium arsenide laser. Invest Ophthalmol13:471 (1974).

20. Gibson, G.L.M. Retinal damage from repeated subthreshold exposuresusing a ruby laser photocoagulator. SAM-TR-70-59, Oct 1970.

21. Gibbons, W.D., and R.G. Allen. Evaluation of retinal damage produced bylong-term exposure to laser radiation. SAM-TR-75-11, Apr 1975.

22. Gibbons, W.D. Retinal burn thresholds for exposure to a frequency-doubled neodymium laser. SAM-TR-73-45, Nov 1973.

23. Hemstreet, H.W., Jr., J.S. Connolly, and D.E. Egbert. Ocular hazards ofpicosecond and repetitive-pulse lasers. Vol. 1: Nd:YAG laser (1064 nm).SAM-TR-78-20, Apr 1978.

24. Cleary, S.F., and P.E. Hamrick. Laser-induced acoustic transients inthe mammalian eye. J Acoust Soc Am 46:1037 (1969).

25. Pitts, D.G., and T.J. Tredici. The effects of ultraviolet radiation onthe eye. Am Indust Hyg Assoc J 32:235 (1971).

26. Ebbers, R.W., and D. Sears. Ocular effects of a 325-nm ultravioletlaser. Am .3 Optom Physiol Opt 52:216 (1975).

27. Zuclich, J.A., and J.S. Connolly. Ocular damage induced by near-ultra-violet laser radiation. Invest Ophthalmol 15:760 (1976).

29. Ham, W.T., Jr., H.A. Mueller, A.I. Goldman, B.E. Newman, L.M. Holland, andT. Kuwabara. Ocular hazard from picosecond pulses of Nd:YAG laserradiation. Science 185:362 (1974).

29. Goldman, A.I., W.T. Ham, Jr., and H.A. Mueller. Ocular damage thresholdsand mechanisms for ultrashort pulses of both visible and infrared laserradiation in the rhesus monkey. Exp Eye Res 24:45 (1977).

30. Taboada, J., and R.W. Ebbers. Ocular tissue damage due to ultrashort1060-nm light pulses from a mode-locked Nd:glass laser. Appl Opt14:1759 (1975).

67

31. Taboada, J., and W.D. Gibbons. Retinal tissue damage induced by singleultrashort 1060-nm laser light pulses. Appl Opt 17:2871 (1978).

32. Boettner, E.A., and J.R. Wolter. Transmission of the ocular media.Invest Ophthalmol 1:776 (1962).

33. Bruckner, A.P., J.M. Schurr, N.B. Martin, and E.L. Chang. Observationof changes induced by picosecond light pulses in suspensions of dipal-mitoyl phosphatidyl choline vesicles. Appl Opt 18:1876 (1979).

34. Bruckner, A.P., J.M. Schurr, and E.L. Chang. Ultrashort laser pulseinduced electromagnetic stress on biological macromolecular systems.SAM-TR-79-3, Nov 1979.

35. Thomas, J.C., S.A. Allison, J.M. Schur, and R.D. Holder. Dynamic lightscattering studies of internal motions in DNA. 11. Clean viral DNA's.Diopoiymers 19:1451 (1980).

36. Bloomfield, V.A., D.M. Crothers, and I. Tinoco, Jr. Physical chemistry ofnucleic acids. New York: Harper and Row, 1974.

37. Lin, S.C., J.C. Thomas, S.A. Allison, and J.'M. Schurr. Dynamic lightscattering studies of internal motions in DNA. III. Evidence fortitratable joints associated with bound polycations, Biopolymers20:209 (1981).

3R. Fangnian, W. Separation of very large DNA molecules by gel electrophoresis.Nucleic Acids Res 5:653 (1978).

39. Letokhov, V.S. On the possibility of selectivw biochemical reactionsinduced by laser radiation. J Photochem 4:185 (1975).

40. Oref, I., and B.S. Rabinovitch. Do highly excitec reactive polyatomicmolecules behave ergodically? Acct Chem Res 12:166 (1979).

41. Jackson, J.D. Classical electrodynamics. New York: John Wiley and Sons,Inc., 1962.

42. Kubo, R. Statistical mechanical theory of irreversible processes. I.General theory and simple applications to maonetic and conductionproblems. J Phys Soc Japan 12:570 (1957).

43. Mathews, J., and R.L. Walker. Mathematical methods of physics. New York:W.A. Benjamin, Inc., 1965.

44. Duguay, M.A., and J.W. Hansen. An ultrafast light gate. Appl Phys

Lett 15:192 (1969).

45. Debye, P. Polar molecules. New York: Dover Publications, Inc., 1929.

46. Bruckner, A.P. Sonp applications of picosecond optical range gating.Proc SPTE 94:41 (1976).

6P

47. Hirrington, R.E. Opticohydrodynamic properties ot high molecularweight DNA. III. The effects of NaCI concentration. Biopolymers17:919 (1978).

48. Voorduow, G., Z. Kam, N. Borochov, and H. Eisenberg. Isolation andphysical studies of the intact supercoiled, the c~onn circular, andthe linear forms of col EI plasmid DNA. Biopolymers 8:171 (1978).

49. Thomas, J.C., S.A. Allison, C.J. Appellof, and J.M. Schurr. Torsiondynamics and depolarization of fluorescence of linear macromolecules.II. Fluorescence polarization anisotropy measurements on a cleanviral 429 DNA. Biophysical Chemistry 12:177 (1980).

50. Kremer, J.M.H., M.W.J. v.d. Esker, C. Pathmamanoharan, and P.H. Wiersma.Vesicles of variable diameter prepared by a modified injection method.Biochemistry 16:3932 (1977).

51. Barenholz, Y., D. Gibbes, B.J. Litman, J. Goll, T.E. Thompson, and F.D.Carlson. A simple method for the preparation of homogeneous phospho-lipid vesicles. Biochemistry 16:2806 (1977).

52. Schurr, J.M. Dynamic light scattering of biopolymers and biocolloids.CRC Critical Reviews of Biochemistry 4:371 (1977).

53. Harris, C. Personal communication. Center for Bioengineering, Universityof Washington, 1979.

54. Finney, D.J. Probit analysis, 2nd ed. New York: Cambridge UniversityPress, 1952.

55. Taboada, J. Personal communication. The standard probit statisticalanalysis of the ocular damage data was conducted in the Laser EffectsBranch, USAF School of Aerospace Medicine, Brooks AFB, Tex., Jan 1980.

56. Shimizu, K., A. Ishimaru, L.O. Reynolds, and A.P. Bruckner. Backscatteringof a picosecond pulse from densely distributed scatterers. Appl Opt18:3484 (1979).

57. Bruckner, A.P. Picosecond light scattering measurements of cataractmicrostructure. Appl Opt 17:3177 (1978).

58. Chen, S.H., W.B. Veldkamp, and C.C. Lai. Simple digital clipped correlatorfor photon correlation spectroscopy. Rev Sci Instrum 46:1356 (1975).

69

APPENDIX A

PICOSECOND LASER IRRADIATION FACILITY

The ultrashort-pulse laser facility assembled for the purpose of irradi-

ating selected macromolecular samples and primate eyes is illustrated in

Figure A-I. It consists of four subsystems: a mode-locked Nd:Glass laser,

a high-speed Pockels cell pulse-switching system, a pulse chronometer and

video detection system, and a pulse-energy measurement system. Each of these

is described below.

Nd:Glass Laser

The mode-locked Nd:Glass laser consists of a water-cooled, Brewster-angled,

Owens-Illinois ED-2 glass rod, 1.3-cm dia x 22.9-cm length, pumped by two EG&G

linear flashlamps in a double elliptical reflector cavity. The resonator is

formed by a flat 99.7% rear reflector (M,) and a 10-m radius 35% output reflec-

tor (M2). Mode-locking is accomplished by a flowing 3-mm-thick dye cell placed

in direct contact with the rear reflector. The dye solution consists of East-

man 9860 dye in dichloroethane at a concentration that results in a small-

signal transmission factor of ,60% at 1060 nm. An iris diaphragm (IDI) is used

to control transverse mode size and purity. By closing it down to an aperture

of 5-mm dia or less, TEM0 0 output can be obtained. In the above configuration

the laser produces a train of n100 horizontally polarized pulses at = 1060 nm,

each of lO-psec duration and >100-MW peak power, spaced at 5.6-nsec intervals

(the round-trip cavity time).

The choice of dye cell geometry is critical to laser performance. We

have experimented with various types of discrete dye cells and with the con-

tacted type, with and without dye circulation, and have found that the circulat-

ing contacted cell produces the most consistent and reproducible pulse trains,

with excellent suppression of satellite pulses. Beam stability and mode purity

are also optimized with this configuration.

At one point in the program we experienced difficulties in proper mode-

locking of our glass laser on account of the poor quality laser-grade

71

C-)Lii

Le)

.V)rj 0

Cl. C L

- c LL- C)*U -r

C)D

-JJ

LnL

oG

im to

IFi 4.W

1.-

L) Un

00

El E - 'to

0.C~ Ul 4iG I k~ .-Li W

_j-

V)iCl- cr a m IlV) uiC) tLU L o,~r ii-

Lii LiJ

LiiI-:

72

iJ. dich loroethane supplied by Eastman Organic Che!nicas. Apparently, high

quantities of residual HCI were present in their then-available lot, No. A6B,

which destroyed the 9860 saturable dye immediately upon mixing. No mode-

locking at all was obtained using this particular lot of solvent. A search

through various chemical distributors turned up some dichloroethane of a differ-

ent lot number (A4), which worked marginally. The dye solution mode-locked

the laser successfully for two or three consecutive days, but then it degraded

and had to be replaced.

Because of these difficulties, for some of the experiments the glass laser

was mode-locked by Eastman 9740 dye in chlorobenzene instead of the usual 9860

dye in dichloroethane. With the 9740 dye the laser produced the same average

total energy per pulse train, but the number of pulses in a train was at most

only ,30-50 compared to the average q,80-100 obtained with 9860 dye. Operation

of the laser was much more erratic with 9740 than with 9860. Multiple and

spurious pulsing occurred on at least half of all mode-locked shots. Only one

in three or four firings of the laser resulted in mode-locking at all. Further-

more, due to the higher peak power obtained with 9740 dye, some of the optical

components sustained some surface or bulk damage. Because of these difficulties,

the use of 9740 dye was discontinued.

The problem of impure dichloroethane was finally solved by filtering the

solvent through a 4" column of basic alumina powder, following the suggestion

of W. Robinson at Texas Tech University. The purified solvent was then Filtered

free of residual alumina particles by means of a 0.45-lim Millipore filter.

The entire purification process is slow but very effective. Excellent mode-

locking has been achieved using the purified dichloroethane as the solvent for

the 9860 dye. Only half the usual concentration of dye is required, and the

solution remains stable for weeks at a time. Mode-locking reproducibility has

improved considprablyalso. Spurious or multiple pulsing occurs no more than

once in 20-30 shots. Although power output is somewhat lower than in the past

(150-200 MW/cm 2 compared to 200-300 'IWlcm2), the pulse duration is less.

Pulses at the beginning of the pulse train have been measured to have durations

in the 6-10-psec range, whereas previously they were in the 10-15-psec range.

73

Lckels Cell Pulse-Switching System

To permit irradiating the macromolecular samples or primate eyes with

single ultrashort pulses as well as with entire pulse trains, a provision for

switching out a single pulse from the mode-locked train has been incorporated.

A high-speed electro-optical shutter is used, consisting of two crossed thin--

film polarizer pairs (TFP) on either side of a high-speed Lasermetrics 1071-FV

Pockels cell, which is switched by a q,7-kV pulse provided by a Lasermetrics

type 8601 avalanche-transistor Krytron-triggered Blumlein pulser. The Pockels

cell is connected to a 50-P terminator via a 30.5-m (100-ft) length of RG-8/U

coaxial cable. The thin-film polarizers are at Brewster's angle (56.50) and

are stacked in pairs to yield a polarization ratio of %2.8x10-5 for each pair.

The output of the laser is horizontally polarized. For convenience the

polarization vector is rotated into the vertical plane by a half-wave retarder

plate (A/2). The first thin-film polarizer rejects any residual horizontal

polarization component. After passing through the inactive Pockels cell, the

first pulse in the train is totally reflected by the second polarizer stack to

an ITT FW 4014 biplanar vacuum photodiode, whose output is displayed on a

Tektronix type 519 oscilloscope.

The first laser pulse triggers the oscilloscope. Approximately 45 nsec

later a fast-rising step-function voltage pulse appears at the "+ Gate" output

of the oscilloscope. This signal is used to trigger the high-voltage pulser

which activates the Pockels cell. The + Gate output delay can be continuously

varied from 45 to 80 nsec, thus permitting precise timing of the pulse deliv-

ered to the Pockels cell. The pulser itself has an additional variable-delay

control, which can be set for pulse delays of u100-400 nsec relative to the

triggering signal if desired. The direct mode, which affords a shorter delay

of -,35 nsec, has been used in our experiments. The laser pulse that happens

to pass through the Pockels cell while it is switched on has its polarization

rotated into the horizontal plane and thus passes through the second polarizcr

uvirpeded. The remaining pulses in the train arrive in the Pockels cell aftey

the switching pulse and hence are totally reflected by the second polarizer.

We have been able to switch out clean single pulses with up to 90 Y effi-

ciency, with a good degree of reproducibility. Generally, about 3 out of 5

shots are successful (i.e., >70 throughput, with no measurable bleed-through

74

of adjacent pulses), provided the mode-locked pulse train is devoid of any

spurious or satellite pulses which could cause premature triggering.

Pulse Chronometer System

Temporal width measurement of the selected ultrashort pulse is carried

out by means of a picosecond streak shutter that we have used extensively inpast work (34,46,56,57). The selected single pulse generates 2nd harmonic

light (530 nm) in a KDP crystal tuned to yield an SHG conversion efficiency

of about 1%. The superimposed infrared (IR) and green pulses are separated

at the dichroic beamsplitter (DBS) (Fig. A-i). For the macromolecular studies

a 50% beamsplitter (BS1 ) directs half the IR pulse energy to the ultrafast

streak shutter. (For the primate experiments all the IR energy is directed

into the streak shutter, since the irradiation of the eye is carried out with

the 2nd or 4th harmonics). The polarization of this pulse is rotated into the

vertical plane by a half-wave retarder ( /2). Lenses L and L2 reduce the

beam diameter by a factor of four. The pulse then traverses a quartz cell

filled with carbon disulfide (CS2 ). This cell is located between two high-

quality crossed polarizers (PI, P2), whose polarization axes are inclined at

45" with respect to the polarization of the IR pulse. These three components

constitute the ultrafast streak shutter. As it travels through the CS2, the

IR pulse induces a narrow zone of birefringence in its immediate vicinity (46).

To an observer viewing the shutter at right angles to the IR path, the effect

is that of a narrow "slit" moving across the line of sight at the speed of

light in CS2 (1.84x1010 cm/sec). The shutter thus produces a streak record

of light pulses incident at right angles to the IR path.

The 530-nm pulse split off at the dichroic beamsplitter is directed

toward the shutter by a right-angle prism (PR) and expanded horizontally by a

pair of cylindrical lenses (L3, L4) to illuminate the entire length of the

shutter, where it is sampled by the IR gating pulse. In the case of the pri-

mate experiments, only a portion of the 530-nm light is directed to the shut-

ter, by means of a pellicle beamsplitter. The signal exiting from the shutter

is a cross-correlation between the gating and green pulses (57). If the depth

of the IR gating pulse is small, the transverse dimension of the transmitted

green pulse is essentially the same as the geometrical pulse length of the IR

pulse in air.

75

The pulses gated by the shutter are detected and processed by the video

detection and display system (VDDS) shown in Figure A-2 (46,57). The shuttr

output is imaged by a Telemation TMC-1100 CCTV camera equipped with an RCA

4532A silicon vidicon tube. The video signal is processed by the video dis-

play control unit (VDC, built in-house) and displayed on an RCA CCTV monitor.

Superimposed on the display is a bright rectangular frame generated by the

VDC. The frame height can be varied from 1 to 64 TV lines, and its width

from an equivalent of 64 lines to full screen width. The intensity profiles

of the TV lines within the frame are displayed on an oscilloscope. The VDDS

is operated in the single-shot mode, wherein only a single sweep of the vidi-

con and display oscilloscope occurs. In this case only half the field of TV

lines is swept; i.e., only the odd- or even-numbered lines. This avoids

charge leakage from the transient image on the vidicon in the time between

sweeps of the odd and even fields. Thus, in this mode up to 32 alternate

lines can be examined. A trigger output pulse from the VDC fires the laser

at the start of the vidicon sweep.

In our studies we have used a frame height of one line and positioned it

to provide a horizontal slice through the center of the cross-correlation

pulse image. The oscilloscope display is thus a plot of pulse intensity as a

function of line. All pulse durations quoted are measured full width at

half-maximum and are deconvoluted for the effect of the finite thickness of

the birefringent zone in the shutter medium.

Pulse Enerqy Measurement

The IR laser pulse energy is monitored by a laser Precision RkP-331

pyroelectric energy probe and RkP-3230 digital display unit. This system was

calibrated by the manufacturer using standards traceable to the National

Bureau of Standards. A 20-cm-f.l. lens placed 10 cm in front of the probe

focuses the incident IR pulse to <3-mm diameter for acceptance by the probe

aperture. The incident pulse is sampled by means of an uncoated glass beam-

splitter (BS2 in Fig. A-I).

76

. oU)~2

z00

'-t-4E-4 -4 z

0 0

0 0 .

>44

-o

'64

00

000

C)

.,__ __ __ _o

0 4J0)

E-4-

b-44

0E-4

ozz

E-4-

le

4-4 4

77

h . . l. . . .. J . . . . . I I - i i . .. . I~ I n . . . . . . . . . . I . . . . ' . .

APPENDIX B

EXPERIMENTAL PROTOCOL FOR DNA STUDIES

Calf-thymus DNA solutions were carefully prepared at a 1-mg/ml con-

centration in aqueous sodium chloride (NaCl) by dissolving the DNA in a cold

room (5C) for periods of 3-5 days, at a carefully controlled stirring rate

of <I cycle/sec. Some samples included the Ca++ sequestering agent, EDTA;

others did not. In each case both the test and control samples were treated

identically except for irradiation by the Nd:Glass laser. Part of the sample

solution was used to fill the selected test cell for irradiation by the pico-

second laser pulses. An identical cell was filled at the same time to act

as the control sample. The test cell was filled carefully to minimize the

occurrence of air bubbles at the optical windows or in the optical path of

the laser pulse.

Two types of UV-grade fused-quartz test cells were manufactured by Preci-

sion Cells, Inc. for the irradiation experiments. One cell type has a 5-mm

ID and a 2-cm path length. Two filler spouts at either end are provided for

filling and flushing the cell. It is referred to as the "short" cell. The

other type is similar in construction to the short cell but has a 2-mm ID

and a 10-cm length and is referred to as the "long" cell. It is used when

higher incident energy levels are produced by focussing down the incident

beam diameter. In each case the incident laser pulse just fills the bore

of the cell and thus irradiates the entire sample.

Test samples were irradiated by single mode-locked pulses, entire pulse

trains, and successions of several pulse trains. To achieve higher energy

densities, the small-diameter long cell was used. The test and control samples

were then taken out of their respective cells and diluted with I-M NaCl solu-

tion in the ratio of 0.5 ml DNA to 9.5 ml NaCl. The solutions were filtered

through either a 3.0- or 1.25-irm filter and collected in scattering cells

for the dynamic light scattering experiments. Sample handling for the gel

electrophoresis and low-shear viscometry is discussed in the main text.

78

APPENDIX C

DYNAMIC LIGHT SCATTERING FACILITY

The dynamic light scattering apparatus consists of a CW He-Ne laser oper-

ating at 632.8 nm with approximately 50-mW power, the optical detection system,

photon counting and correlating electronics, and a PDP-12 computer for data

analysis and storage. The block diagram of the experimental setup is shown

in Figure C-i. The laser and optical detection systems are mounted on a vibra-

tionally damped table which consistsofa 1360-kg, 3.7-m-long, 83-cm-wide-

flange, steel I-beam sitting upon 16 free-floating springs grouped into four

groups of four. The resonance frequency of the beam with respect to Lile floor

is between 1 and 2 Hz, as per design.

The output of the laser is passed through an optical polarizer and a

10-cm-f.l. lens to focus the laser to its minimum waist in the scattering

cell. The optical detection system consists of a low-noise-selected ITT-FW

130 photomultiplier tube with a mu-metal shield enclosed in a specially con-

structed aluminum housing that contains the photomultiplier dynode bias elec-

tronics. The optical detection system is mounted on a triangular optical

rail rigidly attached to a rotary milling table.

A portion of the scattered light from the sample passes through a 10-cm

lens placed such that the light passing through the lens forms a divergent

cone. This is easily achieved by placing the lens so that the distance be-

tween the focused laser-beam waist in the sample and the lens is less than

the focal length of the lens. A series of apertures between the photomulti-

plier and the lens admit light from only a rather small solid angle. The

actual collection solid angle is ultimately determined by the dimension (0.25

mm) of the photoactive cathode and the divergence angle of the beam, and is

slightly less than the coherence solid angle. (Collecting a greater solid

angle simply includes more independently fluctuating K-vectors, reducing the

apparent signal-to-noise ratio.)

The photomultiplier output is fed to the PAR amplifier-discriminator,

which selects the photoelectron pulses and transfers them to the pulse invert-

er. This in turn converts them to TTL pulses suitable for input to the

79

4.

CLL

C)0

L-)

CA I

#0m

00

4,

to 4-

C/) to

4-)0 S-

0 -0 XW )'

CE: LC

900

digital photon-correlator. The output correlation function is stored in the

memory channels of the Nicolet signal averager, which car be transferred to

the PDP-12 computer via its A-D converter.

The digital photon-correlatoris a 256-channel Chen-type digital clipped

correlator (DCC) (58) constructed in-house. A block diagram of the DCC is

shown in Figure C-2.

81

RUN STOP CLEAR

UNITS TS TSIEXP

1 TO 9x1O6 CLOCKCLOCK CONTROLUNITS OF lOOns CIRCUIT CLIPPING LOCK CIRCUIT MCA ADDRESS 611P

MCA ADDRESSI I RESET

CLIP ---- "' "-

CORRELATOR1 iCLIPCLEA RCLI CPE

DCOULTS DISPLAY

zCIRCUTCERCI DCLIP- CLAIx--- CUT IPA

FACTOR lT I/9 O3 AOIE{ OU S

RCLIPPED COUNTS

ITTL SIGNAL DISPLAYS

ECL TO TTL

CONVERT ER = THUMB-WHEEL SWITCH

CIECL SIGNAL IN

Figure C-?. Block diagram of digital clipped correlator.

16BORD